You are on page 1of 55




PGDM PHARMA 2016-2018
ROLL- 01
Artificial intelligence
The study of mechanical or "formal" reasoning began with philosophers and mathematicians in antiquity.
The study of mathematical logic led directly to Alan Turing's theory of computation, which suggested that a
machine, by shuffling symbols as simple as "0" and "1", could simulate any conceivable act of mathematical
deduction. This insight, that digital computers can simulate any process of formal reasoning, is known as the
Church–Turing thesis. Along with concurrent discoveries in neurology, information theory and cybernetics,
this led researchers to consider the possibility of building an electronic brain. The first work that is now
generally recognized as AI was McCullouch and Pitts' 1943 formal design for Turing-complete "artificial

The field of AI research was born at a workshop at Dartmouth College in 1956. Attendees Allen Newell
(CMU), Herbert Simon (CMU), John McCarthy (MIT), Marvin Minsky (MIT) and Arthur Samuel (IBM) became
the founders and leaders of AI research.

Artificial intelligence (AI, also machine intelligence, MI) is intelligence displayed by machines, in contrast
with the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is
defined as the study of "intelligent agents": any device that perceives its environment and takes actions that
maximize its chance of success at some goal. Colloquially, the term "artificial intelligence" is applied when a
machine mimics "cognitive" functions that humans associate with other human minds, such as "learning"
and "problem solving".

The scope of AI is disputed: as machines become increasingly capable, tasks considered as

requiring "intelligence" are often removed from the definition, a phenomenon known as the AI effect,
leading to the quip "AI is whatever hasn't been done yet."] For instance, optical character recognition is
frequently excluded from "artificial intelligence", having become a routine technology. Capabilities generally
classified as AI as of 2017 include successfully understanding human speech, competing at a high level in
strategic game systems (such as chess and Go), autonomous cars, intelligent routing in content delivery
networks, military simulations, and interpreting complex data, including images and videos.

Artificial intelligence was founded as an academic discipline in 1956, and in the years since has experienced
several waves of optimism, followed by disappointment and the loss of funding (known as an "AI winter"),
followed by new approaches, success and renewed funding. For most of its history, AI research has been
divided into subfields that often fail to communicate with each other. These sub-fields are based on
technical considerations, such as particular goals (e.g. "robotics" or "machine learning"), the use of particular
tools ("logic" or "neural networks"), or deep philosophical differences. Subfields have also been based on
social factors (particular institutions or the work of particular researchers).

The traditional problems (or goals) of AI research include reasoning, knowledge, planning, learning, natural
language processing, perception and the ability to move and manipulate objects. General intelligence is
among the field's long-term goals. Approaches include statistical methods, computational intelligence, and
traditional symbolic AI. Many tools are used in AI, including versions of search and mathematical
optimization, neural networks and methods based on statistics, probability and economics. The AI field
draws upon computer science, mathematics, psychology, linguistics, philosophy, neuroscience, artificial
psychology and many others.

The field was founded on the claim that human intelligence "can be so precisely described that a machine
can be made to simulate it".This raises philosophical arguments about the nature of the mind and the ethics
of creating artificial beings endowed with human-like intelligence, issues which have been explored by myth,
fiction and philosophy since antiquity. Some people also consider AI a danger to humanity if it progresses
unabatedly. Others believe that AI, unlike previous technological revolutions, will create a risk of mass

In the twenty-first century, AI techniques have experienced a resurgence following concurrent advances in
computer power, large amounts of data, and theoretical understanding; and AI techniques have become an
essential part of the technology industry, helping to solve many challenging problems in computer science.

Main articles: History of artificial intelligence and Timeline of artificial intelligence

While thought-capable artificial beings appeared as storytelling devices in antiquity, the idea of actually
trying to build a machine to perform useful reasoning may have begun with Ramon Llull (c. 1300 CE). With
his Calculus ratiocinator, Gottfried Leibniz extended the concept of the calculating machine (Wilhelm
Schickard engineered the first one around 1623), intending to perform operations on concepts rather than
numbers. Since the 19th century, artificial beings are common in fiction, as in Mary Shelley's Frankenstein or
Karel Čapek's R.U.R. (Rossum's Universal Robots)ss described as "astonishing": computers were learning
checkers strategies (c. 1954) (and by 1959 were reportedly playing better than the average human), solving
word problems in algebra, proving logical theorems (Logic Theorist, first run c. 1956) and speaking English.
By the middle of the 1960s, research in the U.S. was heavily funded by the Department of Defense and
laboratories had been established around the world. AI's founders were optimistic about the future: Herbert
Simon predicted, "machines will be capable, within twenty years, of doing any work a man can do". Marvin
Minsky agreed, writing, "within a generation ... the problem of creating 'artificial intelligence' will
substantially be solved".

They failed to recognize the difficulty of some of the remaining tasks. Progress slowed and in 1974, in
response to the criticism of Sir James Lighthill and ongoing pressure from the US Congress to fund more
productive projects, both the U.S. and British governments cut off exploratory research in AI. The next few
years would later be called an "AI winter", a period when obtaining funding for AI projects was difficult.

In the early 1980s, AI research was revived by the commercial success of expert systems, a form of AI
program that simulated the knowledge and analytical skills of human experts. By 1985 the market for AI had
reached over a billion dollars. At the same time, Japan's fifth generation computer project inspired the U.S
and British governments to restore funding for academic research. However, beginning with the collapse of
the Lisp Machine market in 1987, AI once again fell into disrepute, and a second, longer-lasting hiatus began.

In the late 1990s and early 21st century, AI began to be used for logistics, data mining, medical diagnosis and
other areas. The success was due to increasing computational power (Moore's law), greater emphasis on
solving specific problems, new ties between AI and other fields and a commitment by researchers to
mathematical methods and scientific standards. Deep Blue became the first computer chess-playing system
to beat a reigning world chess champion, Garry Kasparov on 11 May 1997.

Advanced statistical techniques (loosely known as deep learning), access to large amounts of data and faster
computers enabled advances in machine learning and perception. By the mid 2010s, machine learning
applications were used throughout the world. In a Jeopardy! quiz show exhibition match, IBM's question
answering system, Watson, defeated the two greatest Jeopardy champions, Brad Rutter and Ken Jennings,
by a significant margin. The Kinect, which provides a 3D body–motion interface for the Xbox 360 and the
Xbox One use algorithms that emerged from lengthy AI research as do intelligent personal assistants in
smartphones. In March 2016, AlphaGo won 4 out of 5 games of Go in a match with Go champion Lee Sedol,
becoming the first computer Go-playing system to beat a professional Go player without handicaps. In the
2017 Future of Go Summit, AlphaGo won a three-game match with Ke Jie, who at the time continuously held
the world No. 1 ranking for two years. This marked the completion of a significant milestone in the
development of Artificial Intelligence as Go is an extremely complex game, more so than Chess.

According to Bloomberg's Jack Clark, 2015 was a landmark year for artificial intelligence, with the number of
software projects that use AI within Google increased from a "sporadic usage" in 2012 to more than 2,700
projects. Clark also presents factual data indicating that error rates in image processing tasks have fallen
significantly since 2011. He attributes this to an increase in affordable neural networks, due to a rise in cloud
computing infrastructure and to an increase in research tools and datasets. Other cited examples include
Microsoft's development of a Skype system that can automatically translate from one language to another
and Facebook's system that can describe images to blind people.

Al~tract--The past decade has seen significant advances in medical artificial intelligence (MAI), but its role
in medicine and medical education remains limited. The goal for the next decade must be directed towards
maximizing the utility of MAI in the clinic and classroom. Fundamental to achieving this is increasing the
involvement of clinicians in MAI development. MAI developers must move from "pet projects" toward
generalizable tasks meeting recognized clinical needs. Clinical researchers must be made aware of
knowledge engineering, so clinical data bases can be prospectively designed to contribute directly into MAI
"knowledge bases". Closer involvement of MAI scientists with clinicians is also essential to further
understanding of cognitive processes in medical decision-making. Technological advances in user interfaces--
-including voice recognition, natural language processing, enhanced graphics and videodiscs--- must be
rapidly introduced into MAI to increase physician acceptance. Development of expert systems in non-clinical
areas must expand, particularly resource management, e.g. operating room or hospital admission
scheduling. The establishment of MAI laboratories at major medical centers around the country, involving
both clinicians and computer scientists, represents an ideal mechanism for bringing MAI into the
mainstream of medical computing.


The application of artificial intelligence (AI) techniques to medical problems has been a goal of computing
since the development of high-speed digital computers. In the 1970s, medical artificial intelligence (MAI) was
responsible for the highly successful MYCIN program, the very first expert system . In the 1980s, the rapid
proliferation of AI technology, combined with the greater availability of computers in the medical
environment, led to an increasing number of medical expert systems, such as PUFF, CADUCEUS and
INTERNIST . As in the business world, expert systems in medicine validated themselves when compared
against human "experts" . For the first time, AI began to play a role in medical education as well . At present,
AI programs, such as expert system shells, are available for use on personal computers, and MAI is poised to
become a significant factor in the medical computing environment as we head into the 1990s. Conspicuously
absent on the MAI scene, however, is a firm sense of priority and direction for the development of new
applications, and a sound program for their implementation into the mainstream of medicine. It is the goal
of this paper to suggest, from a clinician's viewpoint, directions and priorities for the next decade of MAI
applications, and suggest a mechanism to implement them. Analogous to the introduction of a new drug,
MAI must pass through a series of phases prior to widespread acceptance. In drug evaluations, Phase I is
directed towards the preliminary evaluation of the new drug, establishing its toxicity (cost) and optimal
mode of delivery. Phase I evaluations are often carried out on patients for whom all conventional therapy
has failed, and expectations for dramatic response to the new treatment are low. Phase I testing does not
establish exactly which patients would be most likely to benefit from the new agent. Phase II testing, by
contrast, is designed to identify specific areas where the new treatment will be effective, and to verify in a
larger number of cases the treatment's safety. Patients who are enrolled in these trials often have less
advanced disease than those in Phase I studies, and it is expected that if a significant benefit exists, it will be
identified in this phase. The groups of patients in Phase II trials are carefully selected, so that if only a small
but distinct population benefits from the new treatment, they will not be lost in the "noise" of the
heterogeneous population of non-responders. Once Phase II has identified specific groups most likely to
respond to the new therapy, and has addressed any hitherto unexpected problems, Phase III evaluation
begins. This phase employs the most effective schedule 314 V.K. SONDAK and N. E. SONDAK of
administration in the patients most likely to benefit, in a head-to-head comparison against the current
standard treatment for the same patients. In this way, the relative value of the new treatment compared to
standard therapy is assessed. If the new method proves itself, it is accepted and approved for general use in
the established "indications". In this analogy, MAI has certainly progressed through Phase I testing. We know
that the technology exists to deliver MAI without excessive cost. On the other hand, "dramatic responses"
have been remarkably few. Even MYCIN and other medical expert systems have been more noteworthy as
research projects than as medical advances. There is, at present, no consensus as to which applications are
best suited to eventual application of MAI, nor is there any data to prove that MAI in any context provides
clear-cut benefit over more conventional alternatives. We have not yet begun Phase II testing in earnest.
Following through with a systematic plan of evaluation for MAI applications, like that described for a new
drug, would have several major advantages. It would force researchers to carefully screen and select the
applications tested, to maximize the likelihood of success. At the same time, it would allow collection of data
in carefully defined, homogeneous settings--so that other researchers, both clinicians and computer
scientists, could assess the generalizability of the MAI system being tested to their own problem areas.
"Single-use only" solutions, which work for the application developer but no one else, would be discouraged,
in favor of more global applications. Once there was some level of agreement on the optimal targets for a
given MAI system, then direct comparisons to the existing standard, whether it be human or "unintelligent"
computer, can be carried out. The results of such a comparison would establish for the medical community
the indications and effectiveness of the new MAI solution. In order to identify application areas likely to
respond to MAI techniques, it is imperative that practicing clinicians are involved in the planning stages of
any new effort. Unfortunately, the clinician--the doctor who must ultimately use the technology--is so often
left out of the loop in the critical early phases of development. A broad base of clinical expertise, particularly
including non-computer-literate physicians, is essential to avoid generating a solution that is technically
correct but technologically inaccessible for the average doctor. Development of interdisciplinary computing
laboratories (combining physicians and computing specialists) should be explored to bring the expertise of
the clinic and computer together. On the other hand, clinical researchers must be made aware of the
concepts and capabilities of knowledge engineering, and "knowledge engineers" should become involved in
the design of clinical data bases. In this fashion, the data collected in clinical data bases, especially in
research and academic settings, can be directly incorporated into "knowledge bases". These knowledge
bases can feed forward into the development of expert systems and other MAI applications, and at the same
time feed back to the clinician valuable information regarding the patients he or she is treating. Improved
access to, and more efficient extraction of useful knowledge from, clinical data bases represents a major
area where closer cooperation between physicians and MAI-oriented computer scientists can lead to
significant gains in the 1990s. At the same time, a closer interaction between clinicans and computer
scientists is needed if the oft-stated goal of improving understanding of the cognitive processes involved in
clinical decision-making is to be achieved. Here again, careful selection of the appropriate "model system" is
essential if computer-simulated decision making systems are to provide useful insight into the mind of the
medical expert. Any computerized system of medical diagnosis or decision-making needs to be carefully
evaluated, first for effectiveness in a carefully defined, homogeneous population ("Phase II"), and then in
head-to-head comparison with the gold standard--the physician himself. Most importantly, computerized
decision-aiding systems need to prove themselves not merely as good as a physician, but better, if they are
truly to be accepted by the medical community (as opposed to the AI world). If clear-cut benefits really exist
from using computers as medical decision-making aids, future MAI trials will have to be designed from their
inception in ways that will allow these benefits to be demonstrated. Physician acceptance of MAI in the
1980s has been hampered by relatively crude and cumbersome user interfaces, one of physicians' major
concerns in evaluating new computer technology, The 1990s hold the promise of major improvements of
interface technology. In particular, natural language processing, voice recognition and improved graphics
should be incorporated into future MAI efforts, as these all address major deficiencies in current user New
directions for MAI 315 interfaces. Most important of all may be the development of videodisc technology
and CD-ROM (compact disc-read only memory); in the image-oriented world of medicine, the ability to store
and quickly retrieve large numbers of photographs, X-rays, etc. will become increasingly vital to any medical
computing venture. While MAI projects with direct clinical application, such as diagnostic tools for
appendicitis, seem to capture the most glamor in AI circles, they are not always clinically useful. We already
have physicians who are making the diagnosis of appendicitis on a daily basis, and they are aware of the
limitations of their diagnostic acumen and have learne d to practice within those limitations. Merely
computerizing the diagnostic process does not extend the limitations, hence, it does not directly further
patient care. Other MAI applications, however, may directly benefit physicians and patients in different
ways. In particular, medical resource management is a fertile, yet underexplored, application area. An
increasing level of cost-consciousness, brought about in part by diagnosis-related group payments (DRGs)
and greater third-party payor surveillance, can be expected to pervade medicine throughout the 1990s.
Clinicians and computer scientists together will need to identify applications, such as bed control (admission,
discharge and transfer tracking), operating room and clinic scheduling, and supply ordering, where expert
systems and other MAI techniques can increase cost-effectiveness unobtrusively. Such applications clearly
need to be a priority area for the next decade. Lastly, questions of hardware and software must be
thoughtfully addressed in designing MAI systems for the immediate future. The extreme variety of personal
and mini-computer systems in the present medical computing environment will surely continue over the
short term. This diversity represents both a challenge and an opportunity for MAI designers. Compatibility
issues should be addressed "up-front", preferably at the design stage, so that when a system proves its value
it can be more easily transferred across hardware barriers. Once again, "single-use only" (or single user
only!) projects must be abandoned in favor of more generalizable solutions as a prerequisite for widespread
adoption of MAI. AI IN MEDICAL EDUCATION Intelligent computer-assisted instruction (ICAI) has developed
in parallel with the entire field of MAI. Given the mandate behind an increased role for medical informatics
in medical education, its future growth seems assured. Nonetheless, many of the caveats stated above for
MAI in general apply equally well to ICAI. In fact, problems of user interface and compatibility loom even
larger in education than in actual practice. With ICAI, MAI is being introduced to a population who may be
only marginally computer literate and who, because of monetary constraints, may have access to only a
single type of computer hardware. Education-related applications need to rely more heavily on explanatory
and error-diagnostic modules than standard MAI ones, so the need for close interaction between medical
educator and computer scientist is great. Perhaps the top priority for computer-aided medical education in
the next decade is the careful selection of appropriate target areas. Medical educators, rather than
computer scientists, will need to take the lead in this selection process, aided by careful studies aimed at
identifying deficiencies in the current (non-computerized) medical education process. Even without waiting
for the completion of such studies, several main targets for ICAI can be identified. Clearly, ICAI--cspecially
when coupled with advanced technologies for graphic representation of information--provides a golden
opportunity to provide "hands on" experience to students at all levels of training. As such, ICAI could
potentially revolutionize medical training in the 1990s, by permitting exposure t0 patient simulators before
physicians arc called upon to treat the actual patient. In clinical practice, scenarios such as trauma
management, critical care and cardiac arrest are usually handled by senior level physicians-in-training
(residents), leaving junior residents, interns and students on the sidelines watching. Suddenly (generally on 1
July of the year), junior residents become senior residents and are now faced with the responsibility for
these emergency situations. ICAI must be evaluated for its ability to case this transition by more adequately
preparing physicians-in-training without jeopardizing overall patient care. Another potential role for ICAI is
to compensate for the inevitable variations in experience that graduates of different medical schools and
residency programs have. For example, a resident who 316 V.K. SO~AK and N. E. SO~AK trained in Southern
California may never have seen a patient with frostbite, while another resident from the Northeast probably
has never seen a rattlesnake bite. Yet when the two residents graduate and begin private practice in
Colarado, they may encounter either of these conditions. Within the same city, residents in training at two
different hospitals---one a tertiary referral center largely filled with patients with unusual conditions, the
other an inner-city hospital populated with patients with neglected illnesses and trauma--will have vastly
different experiences. Medical educators and residency program directors must begin to identify such
regional and inter-hospital variations, and work with computer scientists to develop innovative instructional
techniques which bridge these gaps. Medicine is undergoing an evolutionary change with greatly increased
attention being paid to "cost-effective" delivery of care. Yet most students and residents in "teaching"
hospitals have little or no idea of the cost of the tests and services they are ordering. Professorial teaching
rounds often stress the value of diagnostic tests far more than the costs, leading to a paradoxical situation
where students learn a "shotgun" approach---order every test and you will never be caught without the right
one. Such an approach is no longer tenable, given the fiscal realities of the immediate future. ICAI efforts to
date have not seized the opportunity to incorporate monitors of cost-effectiveness into the overall outcome.
Students participating in a computerized patient evaluation should receive feedback as to the total cost of
the diagnostic tests they ordered, not just the end-result in terms of establishing the correct diagnosis. The
computer should identify ways in which the student could have achieved the same end-result in a more cost-
effective way, selecting only those tests critically necessary to make the proper diagnosis. Such an approach
will, in part, require a rethinking of diagnostic strategies by medical educators as well. This would represent
a valuable, perhaps even vital, "spin-off' of the use of computer technology in medical education. Finally,
medical educators must realize that patients themselves need medical education, and hence are legitimate
targets for MAI efforts. The looming crisis of AIDS provides just one recent example of the need for large-
scale public education measures as an integral part of the management of a disease. Computerized
instruction is an ideal way to convey information directed to the specific needs of the patient, particularly
when the subject is one so emotionally charged as sexual behavior and the risk of AIDS transmission. More
conventional subjects of patient education, such as the need for monthly self breast examination and yearly
mammography for women to reduce their risk of breast cancer, are also worthwhile targets for
computerized instructional efforts. Direct measurement of the effectiveness of computer-aided instruction
compared to conventional education, in terms of modifying patient behavior away from high-risk practices
and toward preventative efforts, should be an integral part of the evaluation of any ICAI effort of this type.
As an additional benefit, patient ICAI could potentially represent a "two-way street", with the computer
asking for and recording information about the patient's current health practices and risk factors at the same
time as it dispenses information about modifying them.


Neural networks (also called parallel distributed processors, neuro computing, connectionist models and
artificial neural systems) are one of the fastest growing and most innovative areas of computing. Neural
networks represent an attempt to simulate biological information processing through massively parallel,
highly-interconnected processing systems. Neural networks offer the potential for solving complex, non-
deterministic problems at very high speeds, the ability to recognize complex patterns, and the capability of
rapidly storing and retrieving very large amounts of information. Neuro computing has received considerable
attention from the U.S. Department of Defense in a number of application areas, including data fusion (the
rapid analysis of data from several different and diverse sources; normally from a variety of electromagnetic
sensors), decision assistance, signal processing and intelligence gathering. Neural networks also have a
number of important commercial applications, such as the dynamic solution of routing problems, image and
handwriting recognition, systems modeling, speech generation, robot control and "expertless" expert
systems . Several of these applications are also of interest in medicine. New directions for MAI 317 Neural
networks have a number of other potentially important medical applications, such as modeling the brain and
nervous system functions, speech analysis and synthesis, X-ray and bacterial culture screening (for
recognition of special types of disease patterns), patient monitoring, as control units for prosthetic devices,
automatic diagnostic systems and the dynamic solution of complex allocation and routing problems in drug
dosage administration, hospital resource allocation, and health care service. Despite the current attention
from the media, neural networks are not a new idea. In fact, neural network concepts, like so many
important advances in AI, have their roots in medicine. Neural network research dates back to 1943, when
Warren McCulloch, a physician, and Walter Pitts, a mathematical physicist, suggested that the complex
computations occurring in the brain could be performed by a network of simple binary neurons performing
elementary logical functions. The McCulloch-Pitts (M-P) neuron model had two types of inputs, an excitory
input and an inhibitory input. The neuron summed the inputs and if the excitory inputs were greater than
the inhibitory inputs, the neuron "fired", that is, generated an output. While the model, as stated, could
account for logical processing, it did not show how information was stored or how intelligent behaviors were
learned [10]. In 1949, Hebb postulated that knowledge was stored in the "connections between the
neurons", and that "learning consisted of modifying these connections and altering the excitory and
inhibitory effects of the various inputs". A number of early experiments with M-P-like neuron networks and
Hebbian learning rules showed very interesting and impressive results . Frank Rosenblatt made a major
contribution to neural network research during this period with the development of the perceptron. The
perceptron, an abstract system based on optical nerve structures, provided a simple model which permitted
extensive mathematical analysis of neural networks. Rosenblatt also pioneered the simulation of these
networks on a digital computer. However, Rosenblatt made claims for his perceptrons which aroused the ire
of a number of other researchers in the field of AI. Marvin Minsky and Seymour Pappert of MIT conducted
an in-depth mathematical analysis of the perceptron and Rosenblatt's claims, which culminated in the
publication of their book Perceptrons. Minsky and Pappert proved theoretically that the perceptron model
was very limited and could not handle large classes of realistic problems . The release of Minsky and
Pappert's work, followed by the untimely death of Rosenblatt in a boating accident, had a very dampening
effect on neural network research. However, limited neural network study continued, even without much
support and funding. The current wave of interest in neural networks began in 1982. John Hopfield, a
prominent biophysicist, showed that artificial neural networks were capable of solving constrained
optimizating problems (such as the "Traveling Salesman Problem", where a salesman must visit a number of
cities in a minimum amount of time and without re-visits). He introduced the concept of a global energy
function to characterize that state of the system, and showed that solutions to equations occupy the lowest
possible energy states, and that artificial neural networks would stabilize or "anneal" to these low-energy
states [13]. Since then, a number of new network and modified learning rules have been developed, some of
which have demonstrated surprising capabilities. There are now sophisticated digital computer programs to
simulate neural networks on personal computers, add-on neural network co-processing boards for personal
and mini-computers, high speed connectivity machines that emulate neural networks, and new chips that
simulate artificial neural systems. Despite all the activity and the increased level of research support for
defense applications, there still is a tremendous amount of work to be done. This is particularly true in the
area of the theoretical understanding of the underlying structure of both biological and artificial neural
systems, as well as the practical area of design and development of medical applications. However, MAI
research and development will play a key role as artificial neural systems begin to realize their potential in
the next decade.
The development of the GUI (graphical user interface; familiar to most of us at the "desktop" metaphor of
the Macintosh computer) during the past decade has marked a significant change in the way humans work
with computer systems. The pioneering research of Xerox with SmaUtalk was translated into a commerical
success by Apple with its Macintosh computer. Steven Jobs' 318 V.K. SONDAK and N. E. SONDAK NEXT
computer system features an extrapolation and enhancement of the GUI. IBM's OS/2 Presentation Manager
and its SAA (system application architecture) graphics standards for applications across the IBM product line,
also indicate the computer industry's commitment to GUI-based interfaces. Intelligent graphics systems (IGS)
represent the next logical development in this area. IGS can be defined as the integration of intelligence with
the standard GUI (windows, icons, mouse, pull-down menus and dialog boxes) devices. The impact of IGS on
MAI could be profound. Specifically, IGS has the potential of significantly increasing the user base for MAI by
allowing the entire spectrum of the health care community direct and simple access to computing
equipment. That is, through IGS, the computer can be made available not only to the experienced physician,
nurse or administrator, but to everyone from the low-level clerical assistant volunteer worker to the patient.
There are three aspects of IGS that should be considered in relation to MAI. The first is the development of
new programming paradigms that will make IGS more accessible. Object-oriented programming (OOP) is
generally regarded as a central isue in the future of graphics and, therefore, IGS. OOP concepts, tools and
techniques should be presented in tutorials or introductory programming courses at the university-level for
medical and nursing students. In addition, medical computer applications developers should consider using
OOP-based programming environments. The second aspect also involves physicians and computer
applications developers. It calls for an extended set of design constructs for medical applications. Current
system design approaches call for the logical design of systems based on the specification of input,
processing, output and storage modules. The emphasis in the input and output areas is typically on data
capture and report generation. The effective use of IGS will require a shift in emphasis to information
transfer. That is, designs should focus on how we can better absorb and understand the situation
confronting the physician rather than the more clerical aspects of data collection and disseminvation. This
will require a complete rethinking of many current medical application packages. The last aspect involves
research into the exact mechanisms of graphic information transfer. We are quite knowledgeable about the
basic physiology of vision but still very ignorant about the how, what and why of the interpretation of
graphic symbols. A great deal of work is yet to be done on how people really absorb and understand graphic
information in general, and in particular, for highly stress-intensive situations that often face medical staffs.
This basic knowledge can be of great value to MAI and IGS applications.


Strategies for implementing the integration of MAI into the mainstream of clinical medicine must revolve
around bringing together clinicians (and medical educators) with computer scientists knowledgeable about
AI. In the reality of today's environment, however, any such collaboration must be academically beneficial
for both parties. The Medical Artificial Intelligence Laboratory (MAIL) represents an ideal venue for the
accomplishment of these goals and objectives. An effective MAIL must be a multidimensional construct. That
is, it must provide an environment for the cross-fertilization of interdisciplinary ideas and expertise.
Personnel should be drawn from medical and health care specialists, computer scientists, electronic
engineers, mathematicians, and information systems specialists. Interests should range from basic research
to applied techniques and technology. The hardware and software tools must be just as broadly-based. The
central thrust of the next decade of computer hardware is towards high-end, graphics-oriented
workstations, networked to large mainframe systems and specialized I/O units. This hardware direction
should be reflected in the MAIL. NEXT computers and RISC workstations will most probably be the
computational platforms of choice. However, research in areas like artificial neural systems, vision systems,
and natural language translation can be very machine cycle-intensive so that networking to super-computers
is a definte requirement. One picture may or may not be worth 1000 words, but in terms of bits and bytes, a
single bit-mapped image using EGA color graphics can require 256 KB. Since physicians are conditioned to
visual information transfer, high resolution color displays are absolutely mandatory for medical applications.
Therefore, most MAI applications of the next decade will have a strong graphics orientation and massive
storage and memories will be needed for the MAIL computer system. New directions for MAI 319 The
traditional computer keyboard with its ancient QWERTY key layout, as the major data entry device for the
physician, is also an area that requires rethinking in the MAIL environment. Improved input devices that are
oriented towards collecting data that the physician uses are in order. For example, the physician collects
data on the visual appearance of the patient, and also auditory, thermal and tactile data. We can reasonably
expect that if we had those types of data captured by appropriate input sensors, we could extend the range
of MAI applications. Therefore, the MAIL computer systems should be flexible enough to accept video
images and have sensor-based input systems that can accept audio, thermal and tactile (strain-gage) data.
Effective use of computer systems requires the use of pointing/selection devices like the mouse, rollerball
and joystick. Current versions of these devices available for desktop computers are quite crude. They
operate in low dots per inch ranges. Development of low-cost, reliable, high resolution pointing/selecting
devices is also a precursor for enhanced MAI applications. This is another area for MAIL study.

Medicine and medical education are changing, and they will continue to do so over the next decade--with or
without AI. The challenge facing clinical practitioners, medical educators and computer scientists alike is to
establish goals and priorities that will allow MAI to assume a fundamental and positive role in these changes.
The generic benefits of AI, particularly in applications involving non-computer-literate users, are clear. The
potential for MAI in the next decade is great. To realize this potential will require the close cooperation of
physicians and computer scientists alike. An essential component to maximize physician adoption of MAI will
be stringent evaluation of MAI systems in prospective trials, combined with greater generalizability than has
been evident to date. The analogy to new drug evaluation is clear, and only a logical sequence of design,
development, and critical evaluation of emerging MAI technologies will assure widespread acceptance of
MAI in the 1990s.

Moving away from generalized medicine to personalization

Classical medical practice puts large groups of people in their focus and tries to develop clinical solutions,
drugs or treatment based on the needs of the statistical average person. Disruptive technologies change that
perspective completely. The basis of that transformation is data. Physicians are able to collect a vast amount
of medical information about the individual through cheap genome sequencing, big data analytics, health
sensors, wearables or artificial intelligence. Based on that specific knowledge, medical professionals can move
away from generalistic solutions towards personalization and precision.

As disruptive technologies appear on the stage of healthcare, it becomes possible to get down even more
deeply to the roots of diseases and treatments. The “one-size-fits-all” strategy will definitely start to crumble.
It is the logical result of hundreds of years of medical research and accumulated knowledge. Currently, we
know that everyone has a different genetic code, may react differently to pharmaceutics or may have a
completely opposite reaction to treatment as assumed.
So why should we treat everyone with the same drugs or with the same method? And one of the most efficient
means for precision medicine is artificial intelligence.

The place of artificial intelligence in precision medicine

As the National Institutes of Health (NIH) put it, precision medicine is “an emerging approach for disease
treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for
each person.” To be able to ponder all those individual variations, medical professionals have to gather
incredible amounts of information, and the ability to analyze, store, normalize or trace that data.

Big data analytics is one area where A.I., especially ANI comes into the picture. Within a couple of years, it will
most probably analyze big medical data sets, draw conclusions, find new correlations based on existing
precedences and support the doctor’s job e.g. in decision-making. Several companies recognized already the
immense potential in A.I. for mining medical records (Google Deepmind and IBM Watson), identifying
therapies (Zephyr Health), supporting radiology (Enlitic, Arterys, 3Scan) or genomics (Deep Genomics). My
personal favorite is Atomwise, which uses supercomputers that root out therapies from a database of
molecular structures. In 2015, Atomwise launched a virtual search for safe, existing medicines that could be
redesigned to treat the Ebola virus. They found two drugs predicted by the company’s A.I. technology which
may significantly reduce Ebola infectivity. This analysis, which typically would have taken months or years, was
completed in less than one day.

Medical limitations and ethical issues around A.I.

To avoid over-hyping technology, the medical limitations of present-day A.I. have to be acknowledged. In the
case of image recognition and using machine learning and deep learning algorithms for the purposes of
radiology, there is the risk of feeding the computer not only with thousands of images but underlying bias. For
example, the images tend to originate from one part of the U.S or the framework for conceptualizing the
algorithm itself incorporates the subjective assumptions of the working team. Moreover, the forecasting and
predictive abilities of smart algorithms are anchored in precedences – however, they might be useless in novel
cases of drug side effects or treatment resistance.

Yet, medical as well as technological limitations of A.I. as well as ANI will still be easier to overcome than ethical
and legal issues. Who is to blame if a smart algorithm makes a mistake and does not spot a cancerous nodule
on a lung X-ray? To whom to turn to when A.I. comes up with a false prediction? Who will build in safety
features? What will be the rules and regulations to decide on safety?

Although these burning questions cannot be answered in their entirety today, we have to do some
preparations to be able to keep the human touch at the center of medicine and avert the possibility of A.I.
becoming an existential threat to mankind feared by Elon Musk or Stephen Hawking.

What should stakeholders do to avoid the A.I. apocalypse?

 Set up ethical standards how to use A.I. on the micro and macro levels of the healthcare sector. We need
specific guidelines starting from the smallest units (medical professionals) until the most complex ones
(national-level healthcare systems). The principle of human comes first should stand at the core of these
 A. I. should be implemented cautiously and gradually in order to give time and space for mapping the
potential risks and downsides. Independent bioethical research groups, as well as medical watchdogs,
should monitor the process closely.
 Medical professionals should familiarize with the basic concepts and working methods of A.I. in a
medical setting to get over their potential fears and understand how the technology could help their
work. There are concerns that A.I. will take over plenty of jobs in healthcare, yet, I believe the key is
cooperation. Medical professionals should work together with technology if they want to achieve their
full potential to heal patients.
 Patients should also explore A.I. in detail and how it might change their own everyday lives. It is
important as in a couple of years, kids will probably play with A.I. friends such as the cute, dinosaur-
shaped Cognitoys or learn from virtual reality teachers.
 Companies who develop A.I. solutions should communicate clearly and concisely towards the general
public about the potential risks of utilizing A.I. in medicine. That’s also useful to avoid overhyping
 Decision-makers at healthcare institutions & policy-makers should guide the process of
implementing A.I. in healthcare along the principles and ethical standards they work out with other
industry stakeholders. Moreover, they should push companies towards putting affordable A.I.
solutions on the table and keeping the focus on the patient all the time.

I have no doubts about that A.I. will be the stethoscope of the 21st century and the backbone of precision
medicine. It has the biggest potential to analyze vast amounts of data and offer insights to create personalized

The transformative power of A.I. in healthcare

We have written extensively about the potential of artificial intelligence for redesigning healthcare. How it
could help medical professionals in designing treatment plans and finding the best-suited methods for every
patient. How it could assist repetitive, monotonous tasks, so physicians and nurses can concentrate on their
actual jobs instead of e.g. fighting the tread-wheel of bureaucracy. By what means A.I. could prioritize e-mails
in doctors’ inboxes or keep them up-to-date with the help of finding the latest and most relevant scientific
studies in seconds. How its transformative power makes it as important as the stethoscope, the symbol of
modern medicine, which appeared in the 19th century.

There are already great examples for its use in several hospitals: Google DeepMind launched a partnership
with the UK’s National Health Service to improve the process of delivering care with digital solutions. In June
2017, DeepMind expanded its services – first of all, its data management app, Streams, to another UK hospital.
IBM Watson is used at the Alder Hey Children’s Hospital as part of a science and technology facilities council
project being run by the Hartree Centre. We asked the first experiences of physicians with the technology and
gave an overview of the ever-expanding line of companies who are extensively investing in the development
of the technology recognizing its transformative capability in healthcare.

However, the question we always have to face is how we translate the vast potential of artificial intelligence
into everyday life. After the very first step – getting to know the most possible about A.I. in healthcare -, we
should get a clearer picture of the obstacles.

2) Medical limitations
To avoid over-hyping the technology, the medical limitations of present-day ANI also have to be
acknowledged. In the case of image recognition and using machine learning and deep learning algorithms for
the purposes of radiology, there is the risk of feeding the computer not only with thousands of images but
also underlying bias.
For example, the images tend to originate from one part of the U.S. or the framework for conceptualizing the
algorithm itself incorporates the subjective assumptions of the working team. Moreover, the forecasting and
predictive abilities of smart algorithms are anchored in previous cases – however, they might be useless in
novel cases of drug side-effects or treatment resistance.

On the other hand, streamlining and standardizing medical records in such a way that algorithms can make
sense of them mean another huge limitation in introducing ANI to hospital departments for doing
administrative tasks. There are many hospitals where doctors still scribble their notes on patients’ files. How
should the computer make sense of such notes if even the person who wrote that cannot read it two weeks

3) Ethical challenges
Yet, medical as well as technological limitations of A.I. as well as ANI will still be easier to overcome than ethical
and legal issues. Who is to blame if a smart algorithm makes a mistake and does not spot a cancerous nodule
on a lung X-ray? To whom could someone turn when A.I. comes up with a false prediction? Who will build in
safety features so A.I. will not turn on humans? What will be the rules and regulations to decide on safety?

These complex ethical and legal questions should be answered if we want to reach the stage of AGI safely and
securely. Moreover, ANI and at a certain point, AGI, should be implemented cautiously and gradually in order
to give time and space for mapping the potential risks and downsides. Independent bioethical research groups,
as well as medical watchdogs, should monitor the process closely. This is exactly what the Open AI
Foundation does on a broader scale. It is a non-profit A.I. research company, discovering and enacting the
path to safe artificial general intelligence. Their work is invaluable, as they are doing long-term research, and
may help in setting up ethical standards on how to use A.I. on micro and macro levels. Perhaps also in the
healthcare sector.

4) Better regulations
The FDA approved the first cloud-based deep learning algorithm for cardiac imaging developed by Arterys in
2017; which is a huge step towards the future. However, regulations around artificial intelligence generally lag
behind or are literally non-existent. With the technology gaining ground and appearing in hospitals within the
next 5-10 years, decision-makers and high-level policy-makers cannot allow themselves not to tackle the issue.

They should rather step ahead of the technological waves and guide the process of implementing A.I. in
healthcare along the principles and ethical standards they work out with other industry stakeholders.
Moreover, they should push companies towards putting affordable A.I. solutions on the table and keeping the
focus on the patient all the time. Governments and policy-makers should also help in setting up standards on
A.I. usage as we need specific guidelines starting from the smallest units (medical professionals) until the most
complex ones (national-level healthcare systems).

5) Misconceptions and overhyping

Overhyping the capabilities of A.I. through marketing tactics and oversimplified media representations does
not help but destroy a healthy image about how A.I. could contribute to healthcare. It also adds to the fog of
confusion and misconceptions which need to be cleared up when we want to implement the technology
successfully into our healthcare systems.
Definitions of machine learning, deep learning, smart algorithms, ANI, AGI or any other terms and concepts
around A.I. need to be treated carefully. The same goes for its impact in healthcare. The story about Facebook
shutting down an A.I. experiment because chatbots developed their own language and how they started
conversing leaving out humans from the process was misrepresented by many news sites from India to Hong
Kong, aggravating fears about A.I. becoming conscious and aiming for destroying the human race. And that’s
just one example out of a swarm of similar articles.

6) Human rejection
Fears about A.I. eradicating humanity go hand-in-hand with exaggerated statements about A.I. coming for the
jobs of medical professionals. Even Stephen Hawking said that the development of full artificial intelligence
could spell the end of the human race. Elon Musk agreed. Moreover, artificial intelligence is said to take the
jobs of radiologists, robots are surpassing the skills of surgeons, or aim to take many jobs in pharma. No
wonder the medical community rejects A.I. Is it not enough for these smart algorithms to take over the world,
are they also coming for our jobs?

The fears around A.I. are understandable as so few of us actually understand how the technology works down
to the detail. And what we don’t understand, we tend to reject. Even more so, if thought-leaders or the media
also tend to treat the issue with exaggeration and extremities. And although it will take time to get accustomed
to the technology, we recommend everyone to be open-minded and familiarize with the concept of using A.I.
in everyday life.

There are various thought leaders who believe that we are experiencing the Fourth Industrial
Revolution,which is characterized by a range of new technologies that are fusing the physical, digital and
biological worlds, impacting all disciplines, economies and industries, and even challenging ideas about what
it means to be human.

I am certain that healthcare will be the lead industrial area of such a revolution and one of the major catalysts
for change is going to be artificial intelligence.

Big Data and Artificial Intelligence Will Revolutionize our Lives

using natural-language processing and then do a limited set of useful things, such as look for a restaurant, get
driving directions, find an open slot for a meeting, or run a simple web search. With the evolution of digital
capacity, more and more data is produced and stored in the digital space. The amount of available digital data
is growing at a mind-blowing speed, doubling every two years. In 2013, it encompassed 4.4 zettabytes,
however by 2020 the digital universe – the data we create and copy annually – will reach 44 zettabytes, or 44
trillion gigabytes (!).

Usually, we make sense of the world around us with the help of rules and processes which build up a system.
The world of Big Data is so huge that we will need artificial intelligence (AI) to be able to keep track of it.

We have not yet reached the state of “real” AI, but it is ready to sneak into our lives without any great
announcement or fanfares – narrow AI is already in our cars, in Google searches, Amazon suggestions and in
many other devices. Apple’s Siri, Microsoft’s Cortana, Google’s OK Google, and Amazon’s Echo services are
nifty in the way that they extract questions from speech

But there is already more to that. A 19-year-old British programmer launched a bot last September which is
successfully helping people to appeal their parking ticket. It is an “AI lawyer” who can sort out what to do with
the received parking ticket based on a few questions. Up until June, the bot has successfully appealed between
160,000 of 250,000 parking tickets in both London and New York, giving it a 64% success rate

Imagine This Efficiency In Healthcare

AI in healthcare and medicine could organize patient routes or treatment plans better, and also provide
physicians with literally all the information they need to make a good decision.
And do not think it is the tale of the distant future. “I have no doubt that sophisticated learning and AI
algorithms will find a place in healthcare over the coming years,” Andy Schuetz, a senior data scientist at Sutter
Health said. “I don’t know if it’s two years or ten — but it’s coming.”

And as winter finally arrived in the sixth season of Game of Thrones, we should be certain that we will gradually
get there. Only by looking at how many companies are interested in AI in healthcare gives the impression that
it is an area with a promising future. Although IBM’s Watson is the big dog in cognitive computing for
healthcare, the race is on and the track is growing increasingly crowded. Dell, Hewlett-Packard, Apple, Hitachi
Data Systems, Luminoso, Alchemy API, Digital Reasoning, Highspot, Lumiata, Sentient Technologies, Enterra,
IPSoft and Next IT – Just to mention a few names.

There are already several great examples of AI in healthcare showing potential implications and possible
future uses that could make us quite optimistic.

However, these solutions will only revolutionize medicine and healthcare if they are available to the average,
mainstream users – and not only to the richest medical institutions (because they are too expensive) or to a
handful of experts (because they are too difficult to use).

Let’s Peek Into The Future

Artificial intelligence already found several areas in healthcare to revolutionize starting from the design of
treatment plans through the assistance in repetitive jobs to medication management or drug drug creation.
And it is only the beginning.

Mining medical records:

The most obvious application of artificial intelligence in healthcare is data management. Collecting it, storing
it, normalizing it, tracing its lineage – it is the first step in revolutionizing the existing healthcare systems.
Recently, the AI research branch of the search giant, Google, launched its Google Deepmind Health project,
which is used to mine the data of medical records in order to provide better and faster health services. The
project is in its initial phase, and at present, they are cooperating with the Moorfields with the Moorfields Eye
Hospital NHS Foundation Trust to improve eye treatment.

Designing treatment plans:

IBM Watson launched its special program for oncologists – and when interviewed one of the professors
working with it – which is able to provide clinicians evidence-based treatment options. Watson for Oncology
has an advanced ability to analyze the meaning and context of structured and unstructured data in clinical
notes and reports that may be critical to selecting a treatment pathway. Then by combining attributes from
the patient’s file with clinical expertise, external research, and data, the program identifies potential treatment
plans for a patient.

IBM launched another algorithm called Medical Sieve. It is an ambitious long-term exploratory project to build
the next generation “cognitive assistant” with analytical, reasoning capabilities and a wide range of clinical
knowledge. Medical Sieve is qualified to assist in clinical decision making in radiology and cardiology. The
“cognitive health assistant” is able to analyze radiology images to spot and detect problems faster and more
reliably. Radiologists in the future should only look at the most complicated cases where human supervision
is useful.

The medical start-up, Enlitic, which also aims to couple deep learning with vast stores of medical data to
advance diagnostics and improve patient outcomes, formulated the perks of deep learning the following
way: “Until recently, diagnostic computer programs were written using a series of predefined assumptions
about disease-specific features. A specialized program had to be designed for each part of the body and only
a limited set of diseases could be identified, preventing their flexibility and scalability. The programs often
oversimplified reality, resulting in poor diagnostic performance, and thus never reached widespread clinical
adoption. In contrast, deep learning can readily handle a broad spectrum of diseases in the entire body, and
all imaging modalities (X-rays, CT scans, etc.)

Getting the most out of in-person and online consultations:

You have a headache, you feel dizzy and you are sure that you have a fever. Your partner tells you that you do
not look great, you should go to the doctor. So, you call the assistant of your GP and ask for an appointment.
It turns out you have to wait two more days to get the chance for a visit. Now, this is what’s not going to
happen with Babylon and its new app. The British subscription, online medical consultation and health
service, Babylon launched an application this year which offers medical AI consultation based on personal
medical history and common medical knowledge. Users report the symptoms of their illness to the app, which
checks them against a database of diseases using speech recognition. After taking into account the patient’s
history and circumstances, Babylon offers an appropriate course of action. The app will also remind patients
to take their medication, and follow up to find out how they’re feeling. Through such solutions, the efficiency
of diagnosing patients can increase by multiple times, while the waiting time in front of doctor’s examining
rooms could drop significantly.

Health assistance and medication management:

Everybody, please welcome the world’s first virtual nurse, Molly developed by the medical start-up
It has a smiling, amiable face coupled with a pleasant voice and its exclusive goal is to help people with
monitoring their condition and treatment. The interface uses machine learning to support patients with
chronic conditions in-between doctor’s visits. It provides proven, customized monitoring and follow-up care,
with a strong focus on chronic diseases.

Also, there is already a solution for monitoring whether patients are taking their medications for real.
The AiCure app supported by The National Institutes of Health uses a smartphone’s webcam and AI to
autonomously confirm that patients are adhering to their prescriptions, or with better terms, supporting
them to make sure they know how to manage their condition. This is very useful for people with serious
medical conditions, for patients who tend to go against the doctor’s advice and participants in clinical trials.
Precision medicine:
Artificial intelligence will have a huge impact on genetics and genomics as well. Deep Genomics aims at
identifying patterns in huge data sets of genetic information and medical records, looking for mutations and
linkages to disease. They are inventing a new generation of computational technologies that can tell doctors
what will happen within a cell when DNA is altered by genetic variation, whether natural or therapeutic.

At the same time, Craig Venter, one of the fathers of the Human Genome Project is working on an algorithm
that could design a patient’s physical characteristics based on their DNA. With his latest enterprise, Human
Longevity, he offers his (mostly affluent) patients complete genome sequencing coupled with a full body
scan and very detailed medical check-up. The whole process enables to spot cancer or vascular diseases in
their very early stage.

Drug creation:
Developing pharmaceuticals through clinical trials take sometimes more than a decade and costs billions of
dollars. Speeding this up and making more cost-effective would have an enormous effect on today’s
healthcare and how innovations reach everyday medicine. Atom wise uses supercomputers that root out
therapies from a database of molecular structures. Last year, Atom wise launched a virtual search for safe,
existing medicines that could be redesigned to treat the Ebola virus. They found two drugs predicted by the
company’s AI technology which may significantly reduce Ebola infectivity. This analysis, which typically would
have taken months or years, was completed in less than one day. “If we can fight back deadly viruses months
or years faster that represents tens of thousands of lives,” said Alexander Levy, COO of Atomwise.“Imagine
how many people might survive the next pandemic because a technology like Atomwise

Another great example of using big data for patient management is Berg Health, a Boston-based biopharma
company, which mines data to find out why some people survive diseases and thus improve current treatment
or create new therapies. They combine AI with the patients’ own biological data to map out the differences
between healthy and disease-friendly environments and help in the discovery and development of drugs,
diagnostics and healthcare applications.

Open AI helping people make healthier choices and decisions:

Did you ever hear the expression, open AI ecosystem? No? Don’t worry, it is rather new and a very fancy
expression for connected AI infrastructures. However, the World Economic Forum named it as one of the top
10 emerging technologies in 2016, so it might be worth getting familiar with it. An open AI ecosystem refers
to the idea that with an unprecedented amount of data available, combined with advances in natural language
processing and social awareness algorithms, applications of AI will become increasingly more useful to

It is especially true in the case of medicine and healthcare. There is so much data to utilize: patient medical
history records, treatment data – and lately information coming from wearable health trackers and sensors.
This huge amount of data could be analyzed in details not only to provide patients who want to be proactive
with better suggestions about lifestyle, but it could also serve healthcare with instructive pieces of
information about how to design healthcare based on the needs and habits of patients.

Analyzing a healthcare system:

97% of healthcare invoices in the Netherlands are digital containing data regarding the treatment, the doctor,
and the hospital. These invoices could be easily retrieved. A local company, Zorgprisma Publiek analyzes the
invoices and uses IBM Watson in the cloud to mine the data. They can tell if a doctor, clinic or hospital makes
mistakes repetitively in treating a certain type of condition in order to help them improve and avoid
unnecessary hospitalizations of patients.

What do we need to make these really happen?

First and foremost, we have to tear down the prejudices and fears regarding artificial intelligence and help the
general population understand how AI could be beneficial and how we can fight its possible dangers. The
biggest fear is that AI will become so sophisticated that it will work better than the human brain and after a
while, it will aim to take control over our lives. Stephen Hawking even said that the development of full artificial
intelligence could spell the end of the human race. Elon Musk agreed.

I do not think that the situation is so gloomy, but I agree with those who stress the need to prepare for the
use of artificial intelligence appropriately. We need the following preparations to avoid the pitfalls of the
utilization of AI:

1. creation of ethical standards which are applicable to and obligatory for the whole healthcare sector
2. the gradual development of AI in order to give some time for mapping of the possible downsides
3. for medical professionals: acquirement of basic knowledge about how AI works in a medical setting in
order to understand how such solutions might help them in their everyday job for patients:
4. getting accustomed to artificial intelligence and discovering its benefits for themselves – e.g. with the help
of Cognitoys which support the cognitive development of small children with the help of AI in a fun and
gentle way or with such services as Siri.
5. for companies developing AI solutions (such as IBM): even more communication towards the general
public about the potential advantages and risks of using AI in medicine.
6. for decision-makers at healthcare institutions: doing all the necessary steps to be able to measure the
success and the effectiveness of the system. It is also important to push companies towards offering
affordable AI-solutions since it is the only way to bring the promise of science fiction into reality and turn
AI into the stethoscope of the 21st century.

If we succeed, huge medical discoveries and treatment breakthroughs will dominate the news not from time
to time, but several times a day. If you ever come across or use a narrow AI system, you will understand my

Bioethical issues on the level of the individual

1) Hacking medical devices
It has already been proven that pacemakers and insulin pumps can be hacked. Security experts have warned
that vulnerabilities could be used to murder patients on a massive scale – sometime soon. The question is –
what can we do to protect wearable devices that are connected to our physiological system from being
hacked and controlled from a distance? Companies developing such technologies should make sure they are
safe and users should be as vigilant as possible when using them.

2) Loss of privacy
We share much more information about ourselves than we think. Check to see what
services and apps you have given permission to access your personal information already. What if, as
augmented reality spreads, all this information will be easily available to someone you just met? Kids who
are born now represent the first generation whose lives are logged in meticulous detail – either by
themselves or well-meaning but clueless relations. While such big data could significantly improve
healthcare, how can we prevent companies and governments from misusing these? What if you ate red
meat and your insurance company immediately raised your insurance rates because you’re not eating
healthy enough?

3) Patients diagnosing themselves at home

Physicians are worried because patients Google their symptoms and treatments, and they might take the
misinformation they find more seriously than what their caregiver tells them. But patients will soon be able
to scan themselves, do blood tests and even genetic analysis on demand with other, unregulated companies
or at home, then use publicly available algorithms to analyse their data. This will open the way for even
more serious cases of misinterpretation, maltreatment or self-medication. Will we able to persuade patients
to turn to doctors with this wealth of data and improve their care, and not just put their trust into
algorithms? If you think this sounds like science fiction, check the finalist of the Nokia Sensing XChallenge,
who have developed just such scanners.

) Healthy people switching to technology

A disruptive technology can provide an unforeseen advantage over others or augment certain human
capabilities to an unprecedented level. As a consequence, what if people start asking their doctors to replace
their healthy limbs for robotic ones because it would let them run faster? What if they start asking for brain
chips to get smarter? If now you can get a new nose or larger breasts, what would prevent you from getting
new muscles or brain implants?

) Bioterrorism
In the wildest futuristic scenarios, tiny nano robots in our bloodstream could detect diseases. These
microscopic robots would send alerts to our smartphones or digital contact lenses before disease could
develop in our body. When most human bodies will contain tiny robots, how can we prevent terrorists from
hacking these devices to gain direct control over our health?

REFERENCES 1. W. Clancy and E. Shortliffe (Eds), Readings in Medical Artificial Intelligence. Addison-
Wesley, Reading, Mass. (1984). 2. P. Harmon and D. King, Expert Systems, pp. 134-154 Wiley, New York
(1984). 3. V. L. Yu, L. M. Fagan, S. M. Wraith, W. J. Clancy, A. C. Scott, J. Hannigan, R~ L. Blum, B. G. Buchanan
and S. N. Cohen, Antimicrobial selection by a computer. A blinded evaluation by infectious diseases experts.
JAMA 242, 1279-1282 (1979). 4. V. K. Sondak and N. E. Sondak, Artificial intelligence and medical education:
an overview and commentary. Proc. 3rd A. artif, lntell, adv Comput. Technol. Conf, Long Beach, California,
pp. 484-489 (1987). 5. R. Schneewiess and V. Cherewatenko, Physicians' attitude toward the uses of
computers in office practice. J. Fam. Prac. 5, 463-465 (1986). 6. R. G. Knapp, M. C. Miller, and J. Levine,
Experience with and attitudes towards computers in medicine of students and clinical faculty members at
one school. J. med. Educ. 62, 344-346 (1987). 7. Executive Council Association of American Medical Colleges,
Evaluation of medical information science in medical education. J. med. Educ. 61, 487-543 (1986). 8. J.
Anderson (Ed.), Neurocomputing: Foundations of Research. MIT Press, Cambridge, Mass. (1988). 9. P. F.
Castelaz, Application of neural network solution to battle management processing functions. 55th milit. Ops.
Res. Soc. Symp., Huntsville, Alabama, May (1987). 10. W. S. McCulloch and W. H. Pitts, A logical calculus of
the ideas imminent in nervous activity. Bull. math. Biophy. 5, 115 (1943). 11. D. O. Hebb, The Organization of
Behavior. Wiley, New York, (1949). 12. M. Minsky and S. Pappert, Perceptrons: An Introduction to
Computational Geometry. MIT Press, Cambridge. 13. J. J. Hopfield, Neural networks and physical systems
with emergent collective computational properties. Proc. Nam. Acad. Sci. 79, 2554-2558 (1982).
1. Coiera, E. (1997). Guide to medical informatics, the internet and telemedicine. Chapman & Hall, Ltd..
2. Jump up^ Power, B. (2015, March 19). Artificial Intelligence Is Almost Ready for Business. Retrieved
3. ^ Jump up to:a b Bloch-Budzier, S. (2016, November 22). NHS using Google technology to treat patients. Retrieved
4. Jump up^ Lorenzetti, L. (2016, April 5). Here's How IBM Watson Health is Transforming the Health Care Industry.
Retrieved from
5. ^ Jump up to:a b CB Insights Artificial Intelligence report. (2016, June 28). Retrieved
6. Jump up^ Lindsay, R. K., Buchanan, B. G., Feigenbaum, E. A., & Lederberg, J. (1993). DENDRAL: a case study of the
first expert system for scientific hypothesis formation. Artificial intelligence, 61(2), 209-261.
7. ^ Jump up to:a b Clancey, W. J., & Shortliffe, E. H. (1984). Readings in medical artificial intelligence: the first decade.
Addison-Wesley Longman Publishing Co., Inc..
8. Jump up^ Bruce, G., Buchanan, B. G., & Shortliffe, E. D. (1984). Rule-based expert systems: the MYCIN experiments
of the Stanford Heuristic Programming Project.
9. Jump up^ Duda, R. O., & Shortliffe, E. H. (1983). Expert systems research. Science, 220(4594), 261-268.
10. Jump up^ Miller, R. A. (1994). Medical diagnostic decision support systems—past, present, and future. Journal of the
American Medical Informatics Association, 1(1), 8-27.
11. Jump up^ Adlassnig KP. A fuzzy logical model of' computer-assisted medical diagnosis. Methods Inf Med. 1980;
12. Jump up^ Reggia, J. A., & Peng, Y. (1987). Modeling diagnostic reasoning: a summary of parsimonious covering
theory. Computer methods and programs in biomedicine, 25(2), 125-134.
13. Jump up^ Baxt, W. G. (1991). Use of an artificial neural network for the diagnosis of myocardial infarction. Annals of
Internal Medicine, 115(11), 843-848.
14. Jump up^ Maclin, P. S., Dempsey, J., Brooks, J., & Rand, J. (1991). Using neural networks to diagnose cancer. Journal
of Medical Systems, 15(1), 11-19.
15. Jump up^ Koomey, J., Berard, S., Sanchez, M., & Wong, H. (2011). Implications of historical trends in the electrical
efficiency of computing. IEEE Annals of the History of Computing, 33(3), 46-54.
16. Jump up^ Dinov, I. D. (2016). Volume and value of big healthcare data. Journal of medical statistics and informatics,
17. Jump up^ Barnes, B., & Dupré, J. (2009). Genomes and what to make of them. University of Chicago Press.
18. Jump up^ Jha, A. K., DesRoches, C. M., Campbell, E. G., Donelan, K., Rao, S. R., Ferris, T. G., ... & Blumenthal, D.
(2009). Use of electronic health records in US hospitals. New England Journal of Medicine, 360(16), 1628-1638.
19. Jump up^ Banko, M., & Brill, E. (2001, July). Scaling to very very large corpora for natural language disambiguation.
In Proceedings of the 39th annual meeting on association for computational linguistics (pp. 26-33). Association for
Computational Linguistics.
20. Jump up^ Dougherty, G. (2009). Digital image processing for medical applications. Cambridge University Press.
21. ^ Jump up to:a b Cohn, Jonathan. "The Robot Will See You Now." The Atlantic, March
22. Jump up^ Spear, Andrew. "From Cancer to Consumer Tech: A Look Inside IBM's Watson Health Strategy." Fortune,
April 5, 2016.
23. Jump up^ Bass, Dina. "Microsoft Develops AI to Help Cancer Doctors Find the Right Treatments." Bloomberg,
September 20, 2016.
24. Jump up^ Knapton, Sarah. "Microsoft Will 'Solve' Cancer within 10 Years by 'Reprogramming' Diseased Cells." The
Telegraph, September 20, 2016.
25. Jump up^ Bloch-Budzier, Sarah. "NHS Using Google Technology to Treat Patients." BBC News, November 22,
26. Jump up^ Lee, Chris Baraniuk, Dave. "Google DeepMind Targets NHS Head and Neck Cancer Treatment." BBC News,
August 31, 2016.
27. Jump up^ Primack, Dan. "Intel Capital Cancels $1 Billion Portfolio Sale." Fortune, May 26,
28. Jump up^ Hernandez, Daniela. "Artificial Intelligence Is Now Telling Doctors How to Treat You." WIRED, June 2,
29. Jump up^ Proffitt, Cas "Top 10 Artificially Intelligent Personal Assistants." Disruptor Daily, Mar 8,
30. ^ Jump up to:a b Felten, Ed. "Preparing for the Future of Artificial Intelligence.", May 3,
31. Jump up^ "The National Artificial Intelligence Research and Development Strategic Plan." Office of Science and
Technology Policy, October 16, 2016.
32. Jump up^ Office of the Press Secretary. "At Cancer Moonshot Summit, Vice President Biden Announces New Actions
to Accelerate Progress Toward Ending Cancer As We Know It.", June 28,
33. Jump up^ Office of the Press Secretary. "President Obama's Precision Medicine Initiative.", January
30, 2015.

1. [1]

Bramer, M.A. and D. Bramer,The Fifth Generation: An Annotated Bibliography, Addison-Wesley, Reading, MA, 1984.Google Scholar

2. [2]

Gevarter, W.B.,Artificial Intelligence, Expert Systems, Computer Vision, and Natural Language Processing, Noyes Publications, Park
Ridge, NJ, 1984.Google Scholar

3. [3]

Glover, F., Future paths for integer programming and links to artificial intelligence,Computers and Operations Research 13(5) (1987)
533–549.Google Scholar

4. [4]

Gonzalez, R.C. and P. Wintz,Digital Image Processing, Addison-Wesley, Reading, MA, 1977.Google Scholar

5. [5]

Papadimitriou, C.H., An algorithm for shortest path motion in three dimensions,Information Processing Letters 20 (1985) 259–
263.Google Scholar

6. [6]

Pratt, W.K.,Digital Image Processing, Wiley, New York, 1978.Google Scholar

7. [7]

Rylko, H.M.,Artificial Intelligence: bibliographic summaries of the select literature, Vol. I, A Report Store Consensus Bibliography,
Lawrence, KS, 1984.Google Scholar

1. [8]
Rylko, H.M.,Artificial Intelligence: bibliographic summaries of the select literature, Vol. II, A Report Store Consensus Bibliography,
Lawrence, KS, 1985.Google Scholar

2. [9]

Wagner, H.,Principles of Management Science, Prentice-Hall, Englewood Cliffs, NJ, 1970.Google Scholar

3. [10]

Andrew, A.M.,Artificial Intelligence, Abacus Press, 1983. (Vol. 3 in the Cybernetics and Systems Series, J. Rose, series ed.)Google

4. [11]

Barr, A. and E.A. Feigenbaum,The Handbook of Artificial Intelligence, Vol. I, William Kaufman, Los Altos, CA, 1981.Google Scholar

5. [12]

Barr, A. and E.A. Feigenbaum,The Handbook of Artificial Intelligence, Vol. II, William Kaufman, Los Altos, CA, 1982.Google Scholar

6. [13]

Bramer, M.A. and D. Bramer,The Fifth Generation: An Annotated Bibliography, Addison-Wesley, Reading, MA, 1984.Google Scholar

7. [14]

Charniak, E. and D.V. McDermott,Introduction to Artificial Intelligence, Addison-Wesley, Reading, MA, 1985.Google Scholar

8. [15]

Clocksin, W.F. and C.S. Mellish,Programming in PROLOG, Springer-Verlag, New York, 1981.Google Scholar

9. [16]

Cohen, P.R. and E.A. Feigenbaum,The Handbook of Artificial Intelligence, Vol. III, William Kaufman, Los Altos, CA, 1982.Google

10. [17]

Dreyfus, H.L. and S.E. Dreyfus,Mind over Machine, MacMillan/The Free Press, New York, 1985.Google Scholar

11. [18]

Feigenbaum, E.A. and J. Feldman, eds.,Computers and Thought, McGraw-Hill, New York, 1963.Google Scholar

12. [19]

Feigenbaum, E.A. and P. McCorduck,The Fifth Generation: Artificial Intelligence and Japan's Computer Challenge to the World,
Addison-Wesley, Reading, MA, 1983.Google Scholar

13. [20]

Fischler, M.A. and O. Firschein,Intelligence: the Eye, the Brain and the Computer, Addison-Wesley, Reading, MA, 1977.Google

14. [21]

Giannesini, F., H. Hanoui, R. Pasero and M. van Caneghem,PROLOG, Addison-Wesley, Reading, MA, 1986.Google Scholar

15. [22]

Nilsson, N.J.,Problem-solving Methods in Artificial Intelligence, McGraw-Hill, New York, 1971.Google Scholar

16. [23]
Nilsson, N.J.,Principles of Artificial Intelligence, Tioga, Palo Alto, CA, 1980.Google Scholar

17. [24]

Rich, E.,Artificial Intelligence, McGraw-Hill, New York, 1983.Google Scholar

18. [25]

Rylko, H.M.,Artificial Intelligence: bibliographic summaries of the select literature, Vol. I, A Report Store Consensus Bibliography,
Lawrence, KS, 1984.Google Scholar

19. [26]

Rylko, H.M.,Artificial Intelligence: bibliographic summaries of the select literature, Vol. II, A Report Store Consensus Bibliography,
Lawrence, KS, 1985.Google Scholar

20. [27]

O'Shea, T. and M. Eisenstadt,Artificial Intelligence: Tools, Techniques and Applications, Harper and Row, New York, 1984.Google

21. [28]

Scientific Datalink, The Scientific Datalink Index to Artificial Intelligence Research (1954–1984) and (1985 supplement), Comtex
Scientific Corporation, New York, (Authors, Titles, Abstracts, Subjects).Google Scholar

22. [29]

Simon, H.A.,Science of the Artificial, M.I.T. Press, Cambridge, MA, 1969.Google Scholar

23. [30]

Winograd, T., R. Davis, S.E. Dreyfus and B. Smith, Expert systems: how far can they go?,Proceedings of IJCAI 9 (1985) 1306–
1309.Google Scholar

24. [31]

Winston, P.H.,Artificial Intelligence, Addison-Wesley, Reading, MA, 1984, 2nd edition.Google Scholar

25. [32]

Winston, P.H. and R.H. Brown,Arificial Intelligence: An MIT Perspective, Vols. 1 and 2, M.I.T. Press, Cambridge, MA, 1979.Google

26. [33]

Blair, C.E., R.G. Jeroslow and J.K. Lowe, Some results and experiments in programming techniques for propositional logic,Computers
and Operations Research 13(5) (1986) 633–645.Google Scholar

27. [34]

Glover, F., Future paths for integer programming and links to artificial intelligence,Computers and Operations Research 13(5) (1986)
533–549.Google Scholar

28. [35]

Grant, T.J., Lessons for O.R. from A.I.: a scheduling case study,Journal of the Operational Research Society 37(1) (1986) 41–
57.Google Scholar

29. [36]

Jeroslow, R.G., An extension of mixed-integer programming models and techniques to some database and Artificial Intelligence
settings, Technical Report, College of Management, Georgia Institute of Technology, March 1987.Google Scholar

30. [37]
Kusiak, A., Applications of operational research models and techniques in flexible manufacturing systems,European Journal of
Operational Research 24(3) (1986) 336–345.Google Scholar

31. [38]

Kusiak, A., Artificial intelligence and operations research in flexible manufacturing systems,INFOR 25(1) (1987) 2–12.Google Scholar

32. [39]

O'Keefe, R.M., Expert systems and operational research — mutual benefits,Journal of the Operational Research Society 36(2) (1985)
125–129.Google Scholar

33. [40]

Paul, R.J. and G.I. Doukidis, Further developments in the use of artificial intelligence techniques which formulate simulatio n
problems,Journal of the Operational Research Society 37(8) (1986) 787–810.Google Scholar

34. [41]

Phelps, R.I., Artificial intelligence, an overview of similarities with O.R.,Journal of the Operational Research Society 37(1) (1986) 13–
20.Google Scholar

35. [42]

Radermacher, F.J., ed.,Proceedings of the International Conference on multi-attribute decision-making via OR-based expert
systems, West Germany, April 1987.Google Scholar

36. [43]

Reggia, J.A., D.S. Nau and Y. Peng, A formal model of diagnostic inference I: problem formulation and decomposition,Information
Sciences 37 (1985) 227–256.Google Scholar

37. [44]

Reggia, J.A., D.S. Nau and Y. Wang, A formal model of diagnostic inference II: algorithmic solution and application,Information
Sciences 37 (1985) 257–285.Google Scholar

38. [45]

Shaw, M.J. and A.B. Whinston, Applications of artificial intelligence to planning and scheduling in flexible manufacturing, in A.
Kusiak ed.,Flexible Manufacturing Systems: Methods and Studies, North-Holland, Amsterdam (1986) 223–242.Google Scholar

39. [46]

Amarel, S., An approach to heuristic problem-solving and theorem proving in the propositional calculus, in: J. Hart and S. Tahasu,
eds.,Systems and Computer Science, University of Toronto Press, Toronto, 1967.Google Scholar

40. [47]

Evans, T.G., A heuristic program to solve geometric analogy problems, in: M. Minsky, ed.,Semantic Information Processing, M.I.T.
Press, Cambridge, MA, 1963.Google Scholar

41. [48]

Kowalski, R., AND/OR graphs, theorem proving graphs, and bidirectional search,Machine Intelligence 7 (1972) 167–194.Google

42. [49]

Kowalski, R.A.,Logic for Problem Solving, Elsevier, New York, 1979.Google Scholar

43. [50]

Laurière, J.L., A language and a program for stating and solving combinatorial problems,Artificial Intelligence 10 (1978) 29–
127.Google Scholar
44. [51]

Levi, G. and F. Sirovich, Generalized AND/OR graphs,Artificial Intelligence 7 (1976) 243–259.Google Scholar

45. [52]

Vanderbrug, G.J., Problem representations and formal properties of heuristic search,Information Sciences 2 (1976) 279–307.Google

46. [53]

Fisher, M.L., Worst-case analysis of heuristic algorithms,Management Science 26 (1980) 1–17.Google Scholar

47. [54]

Graham, R.L., E.L. Lawler, J.K. Lenstra and A.H.G. Rinnooy Kan, Optimization and approximation in deterministic sequencing and
scheduling: a survey,Annals of Discrete Mathematics 5 (1979) 287–326.Google Scholar

48. [55]

Groner, R., M. Groner and W.F. Bischof, eds.,Methods of Heuristics, Lawrence Erlbaum Associates Publishers, 1983.

1. [56]

Hu, T.C. and E.S. Kuh, Theory and concepts of circuit layout, in: T.C. Hu and E.S. Kuh, eds.,VLSI Circuit Layout: Theory and Design,
IEEE Press, New York, 1985.Google Scholar

2. [57]

Johnson, D.S., Approximation algorithms for combinatorial problems,Journal of Computer and System Sciences 9 (1976) 256–
278.Google Scholar

3. [58]

Kusiak, A., ed.,Flexible Manufacturing Systems: Methods and Studies, North-Holland, Amsterdam, 1986.Google Scholar

4. [59]

Lenat, D.B., Discovery in mathematics as heuristic search, Ph.D. Dissertation, Computer Science Department, Carnegie-Mellon
University, 1977.Google Scholar

5. [60]

Lenat, D.B., The nature of heuristics, Computer Science Department, Stanford University, 1980.Google Scholar

6. [61]

Lin, S., Computer solutions of the travelling salesman problem,Bell Systems Technical Journal 44 (1965) 2245–2269.Google Scholar

7. [62]

Minsky, M., Heuristic aspects of the Artificial Intelligence problem, Group Report No. 34-55, Lincoln Laboratory, MIT (ASTIA Doc.
No. AS236885), Lexington, MA, 1956.Google Scholar

8. [63]

Mueller-Merbach, H., Heuristic methods: structures, applications, computational experience, in: R. Cottle and J. Krarup,
eds.,Optimization Methods for Resource Allocation, English Universities Press, London, 1974, 401–416.Google Scholar

9. [64]

Newell, A. and H.A. Simon, Heuristic problem solving: the next advance in Operations Research,Operations Research 6(1)
(1958).Google Scholar

10. [65]
Newell, A. and H.A. Simon,Human Problem Solving, Prentice-Hall, Englewood Cliffs, NJ, 1972.Google Scholar

11. [66]

Pearl, J., Heuristic search theory: a survey of recent results,Proceedings of IJCAI 7, (1981) 24–28.Google Scholar

12. [67]

Pearl, J.,Search and Heuristics, North-Holland, Amsterdam, 1983.Google Scholar

13. [68]

Pearl, J.,Heuristics: Intelligent Search Strategies for Computer Problem Solving, Addison-Wesley, Reading, MA, 1984.Google

14. [69]

Pohl, I., Practical and theoretical considerations in heuristic search algorithms, in: E.W. Elcock and D. Michie, eds.,Machine
Intelligence 8, Wiley, New York, 1977, 55–72.Google Scholar

15. [70]

Slagle, J.R.,Artificial Intelligence: The Heuristic Programming Approach, McGraw-Hill, New York, 1971.Google Scholar

16. [71]

Sussman, G. and R. Stallman, Heuristic techniques in computer-aided circuit analysis,IEEE Transactions on Circuits and Systems,
CAS 22(11) (1975) 857–865.Google Scholar

17. [72]

Thompson, G.L., Heuristic programming: theory and computation, in: J.F. Pierce, ed.,Operations Research and the Design of
Management Information Systems, Techn. Ass. of the Pulp and Paper Industry, New York, 1967.Google Scholar

18. [73]

Zadeh, L.A., Approximate reasoning based on fuzzy logic,Proceedings of IJCAI 6 (1979) 1004–1010.Google Scholar

19. [74]

Deo, N. and C. Pang, Shortest path algorithms: taxonomy and annotations, CS-80-057, Computer Science Department, Washington
State University, 1971.Google Scholar

20. [75]

Gallo, G. and S. Pallottino, Shortest path methods: a unifying approach,Mathematical Programming Studies 26 (1986) 38–64.Google

21. [76]

Horowitz, E. and S. Sahni,Fundamentals of Computer Algorithms, Computer Science Press, Potomac, MD, 1978.Google Scholar

22. [77]

Tarjan, R.E., Depth-first search and linear graph algorithms,SIAM Journal of Computing1 (1972) 146–160.Google Scholar

23. [78]

Bagchi, A. and A. Mahanti, Search algorithms under different kinds of heuristics — a comparative study,Journal of ACM 30(1) (1983)
1–21.Google Scholar

24. [79]

Camerini, P.M. and F. Maffioli, Heuristically guided algorithm fork-parity matroid problems,Discrete Mathematics 21 (1978) 103–
116.Google Scholar
25. [80]

Chakrabarti, P.P., S. Ghose and S.C. De Sarkar, Heuristic search through islands,Artificial Intelligence 29 (1986) 339–347.Google

26. [81]

De Champeaux, B. and L. Sint, An improved bi-directional heuristic search algorithm,Journal of ACM 24 (1977) 177–191.Google

27. [82]

Dechter, R. and J. Pearl, Generalized best first strategies and the optimality ofA*,Journal of ACM 32 (1985) 505–536.Google Scholar

28. [83]

Hart, P.E., N.J. Nilsson and B. Raphael, A formal basis for the heuristic determination of minimum test paths, IEEE Transactions on
Systems and Cybernetics SSC-4(2) (1968) 100–107.Google Scholar

29. [84]

Hart, P.E., N.J. Nilsson and B. Raphael, Correction to a formal basis for the heuristic determination of minimum cost paths,SIGART
Newsletter 37 (1972) 28–29.Google Scholar

30. [85]

Ibaraki, T., The power of dominance relations in branch-and-bound algorithms,Journal of ACM 24(2) (1977) 264–279.Google Scholar

31. [86]

Ibaraki, T., Branch-and-bound procedures and state space representation of combinatorial optimization problems,Information and
Control 36 (1978 a) 1–27.Google Scholar

32. [87]

Ibaraki, T.,M-depth search in branch-and-bound algorithms,International Journal on Computers and Information Science 7 (1978 b)
315–373.Google Scholar

33. [88]

Kernighan, B.W. and S. Lin, An efficient heuristic procedure for partitioning graphs,Bell Systems Technical Journal 29 (February
1970) 291–307.Google Scholar

34. [89]

Korf, R.E., Depth-first iterative-deepening: an optimal admissible tree search,Artificial Intelligence 27 (1985) 97–109.Google Scholar

35. [90]

Krishnamurthy, B., An improved min-cut algorithm for partitioning VLSI networks,IEEE Transactions on Computers C-33(5) (May
1984) 438–446.Google Scholar

36. [91]

Laurierè, J.L., A language and a program for stating and solving combinatorial problems,Artificial Intelligence 10 (1978) 29–
127.Google Scholar

37. [92]

Lawler, E.L. and D.E. Wood, Branch-and-bound methods: a survey,Operations Research14 (1966) 699–719.Google Scholar

38. [93]

Lin, S., Computer solutions of the travelling salesman problem,Bell Systems Technical Journal 44 (1965) 2245–2269.Google Scholar
39. [94]

Martelli, A., On the complexity of admissible search heuristics,Artificial Intelligence 8 (1977) 1–13.Google Scholar

40. [95]

Mérő, L., Some remarks on heuristic search algorithms,Proceedings IJCAI 7 (1981) 572–574.Google Scholar

41. [96]

Mitten, L.G., Branch-and-bound methods: general formulation and properties,Operations Research 18 (1970) 23–34.Google Scholar

42. [97]

Montanari, U., Heuristically guided search and chromosome matching,Artificial Intelligence 1 (1970) 227–245.Google Scholar

43. [98]

Morin, T.L. and R.E. Marsten, Branch and bound strategies for dynamic programming,Operations Research 24 (1976) 611–627.

1. [99]

Nemhauser, G.L., A generalized permanent label setting algorithm for the shortest path between specified nodes,Journal of
Mathematical Analysis and Applications 38 (1972) 328–334.Google Scholar

2. [100]

Pearl, J., Knowledge versus search: a quantitative analysis usingA*,Artificial Intelligence20 (1983) 1–13.Google Scholar

3. [101]

Pearl, J.,Heuristics: Intelligent Search Strategies for Computer Problem Solving, Addison-Wesley, Reading, MA, 1984.Google

4. [102]

Pohl, I., First results on the effect of error in heuristic search, in: B. Meltger and D. Michie, eds., Machine Intelligence 5, Elsevier, New
York, 1970, 219–236.Google Scholar

5. [103]

Amarel, S., An approach to heuristic problem-solving and theorem proving in the propositional calculus, in: J. Hart and S. Tahasu,
eds.,Systems and Computer Science, University of Toronto Press, Toronto, 1967.Google Scholar

6. [104]

Bagchi, A. and A. Mahanti, AND/OR graph heuristic search methods,Journal of ACM 32 (1985) 1–27.Google Scholar

7. [105]

Chang, C.L. and J.R. Slagle, An admissible and optimal algorithm for searching AND/OR graphs,Artificial Intelligence 2 (1971) 117–
128.Google Scholar

8. [106]

Dechter, R. and J. Pearl, Generalized best first search strategies and the optimality ofA*,Journal of ACM 32 (1985) 505–536.Google

9. [107]

Hinxmann, A.I., Problem reduction and the two-dimensional trim-loss problem, in:Proceedings Summer Conference on Artificial
Intelligence and Simulation of Behaviour, Department of Artificial Intelligence, University of Edinburgh, Scotland, 1976.Google

10. [108]
Kowalski, R., AND/OR graphs, theorem proving graphs, and bidirectional search,Machine Intelligence 7 (1972) 167–194.Google

11. [109]

Kumar, V. and L.N. Kanal, A general branch-and-bound formulation for understanding and synthesizing AND/OR tree search
procedures,Artificial Intelligence 21 (1983) 179–198.Google Scholar

12. [110]

Martelli, A. and U. Montanari, Additive AND/OR graphs,Proceedings IJCAI 3(1973) 1–11.Google Scholar

13. [111]

Martelli, A. and U. Montanari, Optimizing decision trees through heuristically guided search,Communications of ACM 21 (1978)
1025–1039.Google Scholar

14. [112]

Nilsson, N.J., Searching problem solving and game-playing trees for minimal cost solutions, in: A.J.H. Morrell, ed.,Information
Processing 68, Vol. 2, North-Holland, Amsterdam, 1969, 1556–1562.Google Scholar

15. [113]

Raiffa, H.,Decision Analysis, Addison-Wesley, Reading, MA, 1968.Google Scholar

16. [114]

Vanderbrug, G.J. and J. Minker, State-space, problem-reduction and theorem proving — some relationships,Communications of
ACM 18 (1975) 107–115.Google Scholar

17. [115]

Bagchi, A. and A. Mahanti, Search algorithms under different kinds of heuristics — a comparative study,Journal of ACM 30(1) (1983)
1–21.Google Scholar

18. [116]

Dechter, R. and J. Pearl, Generalized best first strategies and the optimality ofA*,Journal of ACM 32 (1985) 505–536.Google Scholar

19. [117]

Gaschnig, J., Performance measurement and analysis of certain search heuristics, Ph.D. Dissertation, Computer Science Department,
Carnegie-Mellon University, 1979.Google Scholar

20. [118]

Gelperin, D., On the optimality ofA*,Artificial Intelligence 8(1) (1977) 69–76.Google Scholar

21. [119]

Harris, L.R., The heuristic search under conditions of error,Artificial Intelligence 5(3) (1974) 217–134.Google Scholar

22. [120]

Hart, P.E., N.J. Nilsson and B. Raphael, A formal basis for the heuristic determination of minimum test paths, IEEE Transactions on
Systems Science and Cybernetics SSC-4(2) (1968) 100–107.Google Scholar

23. [121]

Huyn, N., R. Dechter and J. Pearl, Probabilistic analysis of the complexity ofA*,Artificial Intelligence 15 (1980) 241–254.Google

24. [122]
Karp, R.M., The probabilistic analysis of some combinatorial search algorithms, in: J.F. Traub, ed.,Algorithms and Complexity: New
Directions and Recent Results, Academic Press, New York, 1976, 1–19.Google Scholar

25. [123]

Knuth, D.E., Estimating the efficiency of backtrack programs,Mathematics of Computation 24 (1975) 121–136.Google Scholar

26. [124]

Munyer, J., Some results on the complexity of heuristic search in graphs, Technical Report HP-76-2, Information Sciences
Department, University of California, Santa Cruz, 1976.Google Scholar

27. [125]

Pearl, J., Knowledge versus search: a quantitative analysis usingA*,Artificial Intelligence20 (1983) 1–13.Google Scholar

28. [126]

Pohl I., First results on the effect of error in heuristic search, in: B. Meltger and D. Michie, eds.,Machine Intelligence 5, Elsevier, New
York, 1970, 219–236.Google Scholar

29. [127]

Vanderbrug, G.J., Problem representations and formal properties of heuristic search,Information Sciences 2 (1976) 279–307.Google

30. [128]

Banerji, R.B.,Artificial Intelligence: A Theoretical Approach, North-Holland, Amsterdam, 1980.Google Scholar

31. [129]

Ernst, G.W. and M.M. Goldstein, Mechanical discovery of classes of problem-solving strategies,Journal of ACM 29(1) (1982) 1–
23.Google Scholar

32. [130]

Gaschnig, J., A problem similarity approach to devising heuristics: first results,Proceedings IJCAI 6 (1979) 301–307.Google Scholar

33. [131]

Geoffrion, A.M., Elements of large scale mathematical programming. Part I: Concepts, in: A.M. Geoffrion, ed.,Perspectives in
Optimization, Addison-Wesley, Reading, MA, 1972, 25–48.Google Scholar

34. [132]

Guida, G. and M. Somalvico, A method for computing heuristics in problem solving,Information Sciences 19 (1979) 251–259.Google

35. [133]

Kibler, D., Natural generation of admissable heuristics, Department of Computer Science, University of California at Irvine,
1982.Google Scholar

36. [134]

Lenat, D.B., EURISKO: A program that learns new heuristics and domain concepts. The nature of heuristics III: program design and
results,Artificial Intelligence 21(1–2) (1983) 61–98.Google Scholar

37. [135]

Maffioli, F., M.G. Speranza and C. Vercellis, Randomized algorithms: an annotated bibliography,Annals of Operations Research 1
(1984) 331–345.Google Scholar

38. [136]
Mitchell, T.M., R. Utgoff, B. Nudel and R.B. Banerji, Learning problem solving heuristics through practice, Proceedings of IJCAI
7 (1981) 127–134.Google Scholar

39. [137]

Pearl, J., On the discovery and generation of certain heuristics,The UCLA Computer Science Quarterly 10(2) (1982) 121–132,
Reprinted inAI Magazine (Winter/Spring 1983) 23–33.Google Scholar

40. [138]

Valtorta, M., A result on the computational complexity of heuristic estimates for theA* algorithm, Department of Conputer Science,
Duke University, Durham, NC, 1981.Google Scholar

41. [139]

Waterman, D., Generalization of learning techniques for automating the learning of heuristics,Artificial Intelligence 1 (1970) 121–
170.Google Scholar

42. [140]

Bacharach, M.,Economics and the Theory of Games, McMillan, London, 1976.Google Scholar

43. [141]

Berge, C.,Théorie Générale des jeux à n Personnes, Mem. des Sciences Math. 138, Gauthier-Villars, Paris, 1957.Google Scholar

44. [142]

Berlekamp, E.R., J.H. Conway and R.K. Guy,Winning Ways for your Mathematical Plays, Academic Press, New York, 1982.Google

45. [143]

Brams, S.J.,Game Theory and Politics, The Free Press, New York, 1975.Google Scholar

46. [144]

Gillies, D.B., Solutions to general non-zero-sum games,Annals of Mathematics Studies40, Princeton University Press, Princeton, NJ,
1959, 47–85.Google Scholar

47. [145]

Kaplan, E.L.,Mathematical Programming and Games, Wiley, New York, 1982.Google Scholar

48. [146]

Luce, R.D. and H. Raiffa,Games and Decisions, Wiley, New York, 1957.Google Scholar

49. [147]

Owen, G.,Game Theory, W.B. Saunders, Philadelphia, 1968.Google Scholar

50. [148]

Shapley, L.S. and M. Shubik, Characteristic function, core and stable sets,Game Theory in Economics, Ch. 6, Rand Corporation
Report R-904/6-NSF, 1973.Google Scholar

51. [149]

Von Neumann, J. and O. Morgenstern,Theory of Games and Economic Behavior, Princeton University Press, Princeton, NJ, 1944,
1967, 1983.Google Scholar

52. [150]
Acosta, A., Conditions for pathological search on MIN-MAX games, Master's thesis, Department of Engineering Systems, School of
Engineering and Applied Science, University of California, Los Angeles, 1982.Google Scholar

53. [151]

Beal, D., An analysis of minimax, in: M.R.B. Clarke, ed.,Advances in Computer Chess 2, University Press, Edinburgh, 1980, 103–
109.Google Scholar

54. [152]

Beal, D., Benefits of minimax search, in: M.R.B. Clarke, ed.,Advances in Computer Chess 3, Pergamon, Elmsford, NY, 1982, 17–
24.Google Scholar

55. [153]

Berliner, H.J., TheB* tree search algorithm: a best-first proof procedure,Artificial Intelligence 12(1) (1979) 23–40.Google Scholar

56. [154]

Campbell, M.S. and T.A. Marshland, A comparison of minimax tree search algorithms,Artificial Intelligence 20(4) (1983) 347–
367.Google Scholar

57. [155]

Hart, T.P. and D.J. Edwards, The tree prune (TP) algorithm,M.I.T. Artificial Intelligence Project Memo #30, R.L.E. and Computation
Center, M.I.T., Cambridge, MA, 1961.Google Scholar

58. [156]

Ibaraki, T., Generalization of Alpha-Beta and SSS* search procedures,Artificial Intelligence 29 (1986) 73–117.Google Scholar

59. [157]

Knuth, D.E. and R.W. Moore, An analysis of Alpha-Beta pruning,Artificial Intelligence6(4) (1975) 293–326.Google Scholar

60. [158]

Nau, D.S., An investigation of the causes of pathology in games, Technical Report TR-999, Computer Science Department, University
of Maryland, 1981.Google Scholar

61. [159]

Nau, D.S., Pathology on game trees: a summary of results,Proceedings 1st National Conference on Artificial Intelligence (1980) 102–
104.Google Scholar

62. [160]

Nau, D.S., Pathology on game trees revisited and an alternative to minimaxing,Artificial Intelligence 21(1–2) (1983) 221–244.Google

63. [161]

Newell, A., J.C. Shaw and H.A. Simon, Chess-playing programs and the problem of complexity,IBM Journal of Research and
Development 2 (1958) 320–355.Google Scholar

64. [162]

Nilsson, N.J., Searching problem solving and game-playing trees for minimal cost-solutions, in: A.J.H. Morrell, ed.,Information
Processing, North-Holland, Amsterdam, 1969, 1556–1562.Google Scholar

65. [163]

Pearl, J., Asymptotic properties of minimax trees and game searching procedures,Artificial Intelligence 14(2) (1980) 113–138.Google
66. [164]

Pearl, J., On the nature of pathology in game searching,Artificial Intelligence 20(4) (1983) 427–453.Google Scholar

67. [165]

Pearl, J., The solution for the branching factor of the Alpha-Beta pruning algorithm and its optimality,Communications of ACM 25(8)
(1982) 559–564.Google Scholar

68. [166]

Roizen, I. and J. Pearl, A minimax algorithm better than Alpha-Beta? Yes and no,Artificial Intelligence 21(1–2) (1983) 199–
220.Google Scholar

69. [167]

Slagle, J.R. and J.K. Dixon, Experiments with some programs that search game trees,Journal of ACM 16(2) (1969) 189–207.Google

70. [168]

Stockman, G., A minimax algorithm better than Alpha-Beta?Artificial Intelligence 12(2) (1979) 179–196.Google Scholar

71. [169]

Tarsi, M., Optimal search on some game trees,Journal of ACM 30(3) (1983) 389–396.Google Scholar

72. [170]

Verstraete, R., Models for game trees with independent terminal values, Technical Report UCLA-ENG-82-48, Cognitive Systems
Laboratory, University of California, Los Angeles, 1982.Google Scholar

73. [171]

Berliner, H.J., A chronology of computer chess and its literature,Artificial Intelligence10(2) (1978) 201–214.Google Scholar

74. [172]

Kister, J., P. Stein, S. Ulam, W. Walden and M. Wells, Experiments in chess,Journal of ACM 4 (1957) 174–177.Google Scholar

75. [173]

Korf, R.E., A program that learns to solve Rubik's cube,Proceedings of the 2nd AAAI Conference, Pittsburg, PA, 1982, 164–167.Google

76. [174]

Levy, D. and M. Newborn,All About Chess and Computers, Computer Science Press, Rockville, MD, 1982.Google Scholar

77. [175]

Newell, A., J.C. Shaw and H.A. Simon, Chess-playing programs and the problem of complexity,IBM Journal of Research and
Development 2 (1958) 320–355.Google Scholar

78. [176]

Quinlan, J., Knowledge-based system for locating missing high cards in bridge,Proceedings of IJCAI 6 (1979) 705–707.Google Scholar

79. [177]

Samuel, A.L., Some studies in machine learning using the game of checkers,IBM Journal of Research and Development 3 (1959) 211–
229.Google Scholar

80. [178]
Samuel, A.L., Some studies in machine learning using the game of checkers II — recent progress,IBM Journal of Research and
Development 11(6) (1967) 601–617.Google Scholar

81. [179]

Shannon, C.E., Programming a computer for playing chess,Philosophical Magazine 41(7) (1950) 256–275.Google Scholar

82. [180]

Welsh, D.E.,Computer Chess, W.C. Brown, Dubuque, IA, 1984.Google Scholar

83. [181]

Wilkins, D., Using plans in chess,Proceedings of IJCAI 6 (1979) 960–967.


2. 182]

Baudet, G.M., On the branching factor of the Alpha-Beta pruning algorithm,Artificial Intelligence 10(2) (1978) 173–199.Google

3. [183]

Bratko, J. and M. Gams, Error analysis of the minimax principle, in: M.R.B. Clarke, ed.,Advances in Computer Chess 3, Pergamon,
Elmsford, NY, 1982, 1–15.Google Scholar

4. [184]

Fuller, S.H., J. Gaschnig and J.J. Gillogly, An analysis of the Alpha-Beta pruning algorithm, Department of Computer Science Report,
Carnegie-Mellon University, 1973.Google Scholar

5. [185]

Knuth, D.E. and R.W. Moore, An analysis of the Alpha-Beta pruning,Artificial Intelligence6(4) (1975) 293–326.Google Scholar

6. [186]

Newborn, M., The efficiency of the Alpha-Beta search on trees with branch-dependent terminal node scores,Artificial Intelligence 8(2)
(1977) 137–153.Google Scholar

7. [187]

Noe, T., A comparison of the Alpha-Beta and SCOUT algorithms using the game of Kalah, Technical Report UCLA-ENG-80-17,
Cognitive Systems Laboratory, University of California, Los Angeles, 1980.Google Scholar

8. [188]

Pearl, J., The solution for the branching factor of the Alpha-Beta pruning algorithm and its optimality,Communications of ACM 25(8)
(1982) 559–564.Google Scholar

9. [189]

Roizen, I., On the average number of terminal nodes examined by Alpha-Beta, Technical Report UCLA-ENG-81-01, Cognitive Systems
Laboratory, University of California, Los Angeles, 1981.Google Scholar

10. [190]

Andrews, P.B., Theorem proving via general matings,Journal of ACM 28(2) (April 1981) 193–214.Google Scholar

11. [191]

Bibel, W., A comparative study of several proof procedures,Artificial Intelligence 18 (1982) 269–293.Google Scholar

12. [192]
Bibel, W.,Automated Theorem Proving, Vieweg, Braunschweig, 1982.Google Scholar

13. [193]

Biermann, A.W., Automatic programming: a tutorial on formal methodologies,Journal of Symbolic Computation 1 (1985) 119–
142.Google Scholar

14. [194]

Bledsoe, W.W., Non-resolution theorem proving,Artificial Intelligence 9 (1977) 1–35.Google Scholar

15. [195]

Bobrow, D.G., ed., Special issue on non-monotonic logic,Artificial Intelligence 13(1,2) (1980).Google Scholar

16. [196]

Boyer, R.S. and J.S. Moore,A Computational Logic, Academic Press, New York, 1979.Google Scholar

17. [197]

Chang, C.L. and R.C. Lee,Symbolic logic and Mechanical Theorem Proving, Academic Press, New York, 1973.Google Scholar

18. [198]

Cohen, P.R. and E.A. Feigenbaum,The Handbook of Artificial Intelligence, Vol. III, William Kaufman, Los Altos, CA, 1982.Google

19. [199]

Davis, M. and H. Putnam, A computing procedure for quantification theory,Journal of ACM 7(3) (1960) 201–215.Google Scholar

20. [200]

Gelernter, H., Realization of a geometry theorem proving machine, in: E.A. Feigenbaum and J. Feldman, eds., Computers and
Thought, McGraw-Hill, New York, 1963, 134–152.Google Scholar

21. [201]

Kowalski, R.A., A proof procedure using connection graphs,Journal of ACM 22 (1975) 572–595.Google Scholar

22. [202]

Lewis, H.R. and C.H. Papadimitriou,Elements of the Theory of Computation, Prentice-Hall, Englewood Cliffs, NJ, 1981.Google

23. [203]

Loveland, D.,Automated Theorem Proving: A Logical Basis, North-Holland, Amsterdam, 1978.Google Scholar

24. [204]

Lusk, E.L. and R.A. Overbeck, Reasoning about equality,Journal of Automated Reasoning1 (1985) 209–228.Google Scholar

25. [205]

Manna, Z.,Mathematical Theory of Computation, McGraw-Hill, New York, 1974.Google Scholar

26. [206]

Manna, Z. and R. Waldinger,The Logical Basis for Computer Programming, Vol. 1: Deductive Reasoning, Addison-Wesley, Reading,
MA, 1985.Google Scholar

27. [207]
McCharen, J., R. Overbeck and L. Wos, Problems and experiments for and with automated theorem-proving programs,IEEE
Transactions on Computation C-25 (1976).Google Scholar

28. [208]

Newell, A. and H.A. Simon, The logic theory machine,IRE Transactions on Information Theory 2 (1956) 61–79.Google Scholar

29. [209]

Nilsson, N.J.,Principles of Artificial Intelligence, Tioga, Palo Alto, CA, 1980.Google Scholar

30. [210]

Parrelo, B.D., W.C. Kabat and L. Wos, Job-shop scheduling using automated reasoning: A case study of the car sequencing
problem,Journal of Automated Reasoning 2(1) (1986) 1–42.Google Scholar

31. [211]

Paul, E., Equational methods in first order predicate calculus,Journal of Symbolic Computation 1 (1985) 7–29.Google Scholar

32. [212]

Robinson, J.A.,Logic: Form and Function, Edinburgh University Press, Edinburgh, 1979.Google Scholar

33. [213]

Siekmann, J. and G. Wrightson, eds.,Automation of Reasoning 1: Classical Papers on Computational Logic 1957–1966, Symbolic
Computation, Springer-Verlag, Heidelberg, 1983.Google Scholar

34. [214]

Siekmann, J. and G. Wrightson, eds.,Automation of Reasoning 2: Classical Papers on Computational Logic 1967–1970, Symbolic
Computation, Springer-Verlag, Heidelberg, 1983.Google Scholar

35. [215]

Wang, H., Towards mechanical mathematics,IBM Journal of Research and Development4 (1960) 2–22.Google Scholar

36. [216]

Wos, L., R. Overbeck, E. Lusk and J. Boyle,Automated Reasoning: Introduction and Applications, Prentice-Hall, Englewood Cliffs,
NJ, 1984.Google Scholar

37. [217]

Wos, L., F. Pereira, R. Hong, R.S. Boyer, J.S. Moore, W.W. Bledsoe, L.J. Henschen, B.G. Buchanan, G. Wrightson and C. Green, An
overview of automated reasoning and related fields,Journal of Automated Reasoning 1 (1985) 5–48.Google Scholar

38. [218]

Wu, W.-T., Basic principles of mechanical theorem proving in elementary geometries,Journal of Automated Reasoning 2 (1986) 221–
252.Google Scholar

39. [219]

Freuder, E.C., Synthesizing constraint expressions,Communications of ACM 21(11) 1978.Google Scholar

40. [220]

Gaschnig, J., A general backtrack algorithm that eliminates most redundant tests,Proceedings of the International Conference on
Artificial Intelligence, Cambridge, MA (1977) 457.Google Scholar

41. [221]
Gaschnig, J., Performance measurement and analysis of certain search algorithms, Ph.D. thesis, Department of Computer Science ,
Carnegie-Mellon University (May, 1979).Google Scholar

42. [222]

Haralick, R.M., Scene analysis: homomorphism and arrangements, in: Hanson and Reisman, eds.,Machine Vision, Academic Press,
New York, 1978.Google Scholar

43. [223]

Haralick, R.M. and G.L. Elliott, Increasing tree search efficiency for constraint satisfaction problems, Artificial Intelligence 14 (1980)
263–313.Google Scholar

44. [224]

Haralick, R.M. and J. Kartus, Arrangements, homomorphisms and discrete relaxation,IEEE Transactions on Systems, Man and
Cybernetics SMC-8 (August 1978) 600–612.Google Scholar

45. [225]

Haralick, R.M. and L. Shapiro, The consistent labeling problem: PartI,IEEE Transactions on Pattern Analysis and Machine
Intelligence 1(2) (April 1979) 173–184.Google Scholar

46. [226]

Haralick, R.M. and L. Shapiro, The consistent labeling problem: Part II,IEEE Transactions on Pattern Analysis and Machine
Intelligence 2(3) (1980) 193–203.Google Scholar

47. [227]

Haralick, R.M., L.S. Davis, A. Rosenfeld and D.L. Milgram, Reduction operations for constraint satisfaction,Information Sciences 14
(1978) 199–219.Google Scholar

48. [228]

Kowalski, R.A., A proof procedure using connection graphs,Journal of ACM 22 (October 1975) 572–595.Google Scholar

49. [229]

Mackworth, A.K., Consistency in networks of relations,Artificial Intelligence 8(1) (1977) 99–118.Google Scholar

50. [230]

Mackworth, A.K. and E.C. Freuder, The complexity of some polynomial network consistency algorithms for constraint satisfactio n
problems,Artificial Intelligence 25 (1985) 65–74.Google Scholar

51. [231]

McGregor, J.J., Relational consistency algorithms and their applications in finding subgraph and graph isomorphisms, Information
Sciences 19 (1979) 229–250.Google Scholar

52. [232]

Mohr, R. and T.C. Henderson, Arc and path consistency revisisted,Artificial Intelligence28 (1986) 225–233.Google Scholar

53. [233]

Montanari, U., Networks of constraints: fundamental properties and applications to picture processing,Information Sciences 7 (1976)
95–132.Google Scholar

54. [234]

Nudel, B., Consistent-labeling problems and their algorithms: expected-complexities and theory-based heuristics,Artificial
Intelligence 21 (1983) 135–178.Google Scholar
55. [235]

Rosenfeld, A., R. Hummel and S. Zucker, Scene labeling by relaxation operations,IEEE Transactions on Systems, Man and
Cybernetics SMC-6 (1976) 420–423.Google Scholar

56. [236]

Shapiro, L.G. and R.M. Haralick, Structural descriptions and inexact matching,IEEE Transactions on Pattern Analysis and Machine
Intelligence 3 (1981) 504–519.Google Scholar

57. [237]

Ullman, J.R., An algorithm for subgraph homomorphisms,Journal of ACM 23(1) (1976) 31–42.Google Scholar

58. [238]

Waltz, D.L., Generating semantic descriptions from drawings of scenes with shadows, MAC AI-TR-271, M.I.T. 1972.Google Scholar

59. [239]

Aspvall, B., M.F. Plass and R.E. Tarjan, A linear-time algorithm for testing the truth of certain quantified Boolean
formulas,Information Processing Letters 8 (1979) 121–123.Google Scholar

60. [240]

Balas, E. and J.N. Hooker, Applying polyhedral methods to logical inference, paper presented at the TIMS/ORSA meeting, New
Orleans, LA, May 1987.Google Scholar

61. [241]

Blair, C.E., R.G. Jeroslow and J.K. Lowe, Some results and experiments in programming techniques for propositi onal logic,Computers
and Operations Research 13(5) (1986) 633–645.Google Scholar

62. [242]

Bugrara, K. and C.A. Brown, On the average case analysis of some satisfiability model problems,Information Sciences 40 (1986) 21–
37.Google Scholar

63. [243]

Chang, C.L. and R.C. Lee,Symbolic Logic and Mechanical Theorem Proving, Academic Press, New York, 1973.Google Scholar

64. [244]

Cook, S.A., The complexity of theorem-proving procedures,Proceedings 3rd ACM Symposium on Theory of Computing (May 1971)
151–158.Google Scholar

65. [245]

Davis, M. and H. Putnam, A computing procedure for quantification theory,Journal of ACM 7 (1960) 201–215.Google Scholar

66. [246]

Desrosiers, J., B. Jaumard and L. Le Bars, Algorithms for the satisfiability problem, GERAD-HEC Research Report, to appear.Google

67. [247]

Franco, J., Probabilistic analysis of the pure literal heuristic for the satisfiability problem,Annals of Operations Research 1 (1984)
273–289.Google Scholar

68. [248]

Franco, J., On the probabilistic performance of algorithms for the satisfiability problem,Information Processing Letters 23 (1986)
103–106.Google Scholar
69. [249]

Franco, J. and M. Paull, Probabilistic analysis of the Davis—Putman procedure for solving the satisfiability problem,Discrete Applied
Mathematics 5 (1983) 77–87.Google Scholar

70. [250]

Gallo, G. and G.P. Urbani, New algorithms for the satisfiability problem, Nota Scientifica 5-87-6, Dipartimen to Informatica,
Universitá di Pisa, March 1987.Google Scholar

71. [251]

Goldberg, A., P.W. Purdom and C.A. Brown, Average time analyses of simplified Davis—Putnam procedures,Information Processing
Letters 15(2, 6) (September 1982) 72–75 (Errata 16 (1983) 213).Google Scholar

72. [252]

Hammer, P.L. and S. Rudeanu,Boolean Methods in Operations Research and Related Areas, Springer-Verlag, Heidelberg,
1968.Google Scholar

73. [253]

Hansen, P. and B. Jaumard, Algorithms for the maximum satisfiability problem, RUTCOR Research Report, Rutgers University,
November 1987.Google Scholar

74. [254]

Hansen, P., B. Jaumard and M. Minoux, A linear-expected time algorithm for deriving all logical conclusions implied by a set of
Boolean inequalities,Mathematical Programming34 (1986) 223–231.Google Scholar

75. [255]

Jeroslow, R.G., Ten lectures on mixed integer model formulations for logic-based decision support, Technical Report, College of
Management, Georgia Institute of Technology, March 1987 (to appear in theAnnals of Discrete Mathematics).Google Scholar

76. [256]

Johnson, D.S., Approximation algorithms for combinatorial problems,Journal of Computer and System Sciences 9 (1976) 256–
278.Google Scholar

77. [257]

Karp, R.M., Reducibility among combinatorial problems, in: R.G. Miller and J.W. Thatcher, eds.,Complexity of Computer
Computations, Plenum press, New York (1972) 85–104.Google Scholar

78. [258]

Lieberherr, K.J., Algorithmic extremal problems in combinatorial optimization,Journal of Algorithms 3 (1982) 225–244.Google

79. [259]

Lieberherr, K.J. and E. Specker, Complexity of partial satisfaction,Journal of ACM 28(2) (1981) 411–421.Google Scholar

80. [260]

Nilsson, N.J., Probabilistic logic,Artificial Intelligence 28 (1986) 71–88.Google Scholar

81. [261]

Petreschi, R. and B. Simeone, A switching algorithm for the solution of quadratic Boolean equations,Information Processing
Letters 11 (1980) 193–198.Google Scholar

82. [262]
Purdom, P.W., Evaluating search methods analytically,Proceedings National Conference on Artificial Intelligence, Pittsburg, PA
(1982) 124–127.Google Scholar

83. [263]

Purdom, P.W., Search rearrangement backtracking and polynomial average time,Artificial Intelligence 21 (1983) 117–133.Google

84. [264]

Purdom, P.W. and C. Brown, The pure literal rule and polynomial average time,SIAM Journal of Computing 14(4) (November 1985)
943–953.Google Scholar

85. [265]

Rudeanu, S.,Boolean Functions and Equations, North-Holland, Amsterdam, 1974.Google Scholar

86. [266]

Schaefer, T.J., The complexity of satisfiability problems,Proceedings 10th Annual ACM Symposium on Theory of Computing (1978)
216–226.Google Scholar

87. [267]

Shostak, R.E., Refutation graphs,Artificial Intelligence 7 (1976) 51–64.Google Scholar

88. [268]

Simeone, B., Consistency of quadratic Boolean equations and the König-Egerváry property for graphs,Annals of Discrete
Mathematics 25 (1985) 281–290.Google Scholar

89. [269]

Tovey, C.A., A simplified NP-complete satisfiability problem,Discrete Applied Mathematics 8 (1984) 85–89.Google Scholar

90. [270]

Arvind, V. and S. Biswas, An O(n 2) algorithm for the satisfiability problem of a subset of propositional sentences in CNF that includes
all Horn sentences,Information Processing Letters 24 (1987) 67–69.Google Scholar

91. [271]

Dowling, W.F. and J.H. Gallier, Linear-time algorithms for testing the satisfiability of propositional Horn formulae,Journal of Logic
Programming 3 (1984) 267–284.Google Scholar

92. [272]

Henschen, L.J., Semantic resolution for Horn sets,IEEE Transactions on Computers25(8) (1976).Google Scholar

93. [273]

Henschen, L. and L. Wos, Unit refutations and Horn sets,Journal of ACM 21(4) (1974) 590–605.Google Scholar

94. [274]

Jaumard, B. and B. Simeone, On the complexity of the maximum satisfiability problem for Horn formulas,Information Processing
Letters 26 (1987/88) 1–4.Google Scholar

95. [275]

Jones, N.D. and W.T. Laaser, Complete problems for deterministic polynomial time,Theoret. Computer Science 3 (1976) 105–
117.Google Scholar

96. [276]
Lewis, H.R., Renaming a set of clauses as a Horn set,Journal of ACM 25(1) (January 1978) 134–135.Google Scholar

97. [277]

Mannila, H. and K. Mehlhorn, A fast algorithm for renaming a set of clauses as a Horn set,Information Processing Letters 21 (1985)
269–272.Google Scholar

98. [278]

Sykora, O., An optimal algorithm for renaming a set of clauses into the Horn set,Computer and Artificial Intelligence 4(1) (1985) 37–
43.Google Scholar

99. [279]

Tärnlund, S.A., Horn clause computability,BIT 17 (1977) 215–216.Google Scholar

100. [280]

Yamasaki, S. and S. Doshita, The satisfiability problem for a class consisting of Horn sentences and some non-Horn sentences in
proportional (sic!) logic,Information and Control 59 (1983) 1–12.Google Scholar

101. [281]

Yamasaki, S., M. Yoshida, S. Doshita and M. Hirata, A new combination for input and unit deductions for Horn
sentences,Information Processing Letters 18 (1984) 209–213.Google Scholar

102. [282]

Dilger, W. and A. Janson, Intelligent backtracking in deduction systems by means of extended unification graphs,Journal of
Automated Reasoning 2 (1986) 43–62.Google Scholar

103. [283]

Eder, E., Properties of substitutions and unifications,Journal of Symbolic Computation 1 (1985) 31–46.Google Scholar

104. [284]

Fages, F. and G. Huet, Unification and matching in equational theories,CAAP 83, Lecture Notes in Computer Science 159, Springer-
Verlag, Heidelberg, 1983.Google Scholar

105. [285]

Martelli, A. and U. Montanari, An efficient unification algorithm,ACM Transactions on Programming Languages of Systems 4(2)
(1982) 258–282.Google Scholar

106. [286]

Morokhovets, M.K., A modified unification procedure,Cybernetics 20(20) 1 (1984) 147–152.Google Scholar

107. [287]

Paterson, M.S. and M.N. Wegman, Linear unification,Journal of Computers and System Sciences 16(1978) 158–167.Google Scholar

108. [288]

Apt, K.R. and M.H. Van Emden, Contributions to the theory of logic programming,Journal of ACM 29(3) (July 1982) 841–862.Google

109. [289]

Balbin, I. and K. Lecot,Logic Programming: A Classified Bibliography, Wildgrass Books, Fitzroy, Australia, 1985.

1. [290]
Clark, K.L., Negation as failure, in: H. Gallaire and J. Minker, eds.,Logic and Databases, Plenum Press, New York, 1978, 293–
322.Google Scholar

2. [291]

Clark, K.L. and S.A. Tärnlund,Logic Programming, Academic Press, New York, 1982.Google Scholar

3. [292]

Clocksin, W.F. and C.S. Mellish,Programming in PROLOG, Springer-Verlag, Heidelberg, 1981.Google Scholar

4. [293]

Coelho, H. and L.M. Pereira, Automated reasoning in geometry theorem proving with PROLOG,Journal of Automated Reasoning 2(4)
(1986) 329–390.Google Scholar

5. [294]

Colmerauer, A., H. Hanoui, P. Russel and R. Pasero, Un système de communication homme-machine en Francais, Groupe de
Recherche en Intelligence Artificielle, Université d'Aix-Marseille, 1973.Google Scholar

6. [295]

Gabbay, D.M. and M.J. Sergot, Negation as inconsistency, Department of Computing Reports, Imperial College of Science and
Technology, London, England, 1984.Google Scholar

7. [296]

Gallaire, H. and J. Minker, eds.,Logic and Databases, Plenum Press, New York, 1978.Google Scholar

8. [297]

Gallier, J.H. and S. Raatz, Logic programming and graph rewriting, Research Report, Department of Computer and Information
Science, University of Pennsylvania, May 1985.Google Scholar

9. [298]

Hayes, P.J., Computation and deduction,Proceedings 2nd Symposium on Mathematical Foundation of Computer Science,
Czechoslovak Academy of Sciences (1973) 105–118.Google Scholar

10. [299]

Hogger, C.J.,Introduction to Logic Programming, APIC Studies in Data Processing 21, Academic Press, New York, 1984.Google

11. [300]

Jaffar, J., J.L. Lassez and J.W. Lloyd, Completeness of the negation as failure rule,Proceedings of IJCAI 8, 1983.Google Scholar

12. [301]

Kowalski, R.A., Predicate logic as a programming language, in: J.L. Rosenfeld, ed.,Information Processing 74, North-Holland,
Amsterdam, 1974, 569–574.Google Scholar

13. [302]

Kowalski, R.A., Algorithm = Logic + Control,Communications of ACM 22(7) (July 1979) 424–436.Google Scholar

14. [303]

Kowalski, R.A.,Logic for Problem Solving, Elsevier, New York, 1979.Google Scholar

15. [304]

Kowalski, R.A. and D. Kuehner, Linear resolution with selection function,Artificial Intelligence 2 (1970) 227–260.Google Scholar
16. [305]

Lloyd, J.W.,Foundations of Logic Programming, Symbolic Computation, Springer-Verlag, Berlin, 1984.Google Scholar

17. [306]

Pereira, L.M., Logic control with logic,Proceedings 1st International Logic Programming Conference, Luminy, France (September,
1982).Google Scholar

18. [307]

Robinson, J.A.,Logic: Form and Function, Edinburgh University Press, Edinburgh, 1979.Google Scholar

19. [308]

Van Emden, M.H. and R.A. Kowalski, The semantics of predicate logic as a programming language,Journal of ACM 23(4) (1976) 733–
742.Google Scholar

20. [309]

Warren, D.H.D., L.M. Pereira and F. Pereira, PROLOG — the language and its implementation compared with LISP,Proceedings of
the Symposium on Artificial Intelligence and Programming Languages (ACM); SIGPLAN Notices 12(8); andSIGART Newsletter 64
(1977) 109–115.Google Scholar

21. [310]

Warren, D.H.D. and M. Van Caneghem,Logic Programming and its Applications, Albex, Norwood, NJ, 1985.Google Scholar

22. [311]

Aho, A.V., C. Beeri and J.D. Ullman, The theory of joins in relational databases,ACM Transactions on Database Systems (4,3)
(September, 1979) 297–314.Google Scholar

23. [312]

Aho, A.V., Y. Sagiv and J.D. Ullman, Equivalence among relational expressions,SIAM Journal on Computing 8(2) (1979) 218–
246.Google Scholar

24. [313]

Asirelli, P., M. De Santis and A. Martelli, Integrity constraints in logic databases,Journal of Logic Programming 3 (1985) 221–
232.Google Scholar

25. [314]

Ausiello, G., A. D'Atri and D. Sacca, Graph algorithms for functional dependency manipulation,Journal of ACM 30(4) (1983) 752–
766.Google Scholar

26. [315]

Chandra, A.K. and D. Harel, Horn clause queries and generalizations,Journal of Logic Programming 1 (1985) 1–15.Google Scholar

27. [316]

Chang, C.L., Deduce 2: Further investigations of deductions in relational databases, in: H. Gallaire and J. Minker, eds., Logic and
Databases, Plenum Press, New York, 1978, 201–236.Google Scholar

28. [317]

Colmerauer, A. and J.F. Pique, About natural logic, in: H. Gallaire, J. Minker and J.M. Nicolas, eds.,Advances in Database Theory,
Vol. 1, Plenum Press, New York, 1981, 343–345.Google Scholar

29. [318]

Delobel, C. and R.G. Casey, Decomposition of a database and the theory of Boolean switching functions, IBM Journal of Research and
Development 17(5) (September 1973) 484–485.Google Scholar
30. [319]

Demolombe, R., Estimation of the number of triples satisfying a query expressed in predicate calculus language, Proceedings 6th
International Conference on Very Large Data Bases, IEEE, Montreal (October 1980) 55–63.Google Scholar

31. [320]

Fagin, R., Horn clauses and database dependencies,Journal of ACM 29(4) (1982) 952–985.Google Scholar

32. [321]

Gallaire, H., Impacts of logic on data bases,Proceedings 7th International Conference on Very Large Data Bases (Cannes, France,
September 9–11), IEEE, New York (1981) 248–259.Google Scholar

33. [322]

Gallaire, H. and J. Minker,Logic and Databases, Plenum Press, New York, 1978.Google Scholar

34. [323]

Gallaire, H., J. Minker and J.M. Nicolas, Logic and databases: a deductive approach,Computing Surveys 16(2) (1984) 153–185.Google

35. [324]

Gardarin, G. and M. Melkanoff, Proving consistency of database transactions,Proceedings 5th International Conference on Very
Large Data Bases (Rio de Janeiro, October 3–5), IEEE, New York (1979) 291–298.Google Scholar

36. [325]

Gray, P.,Logic, Algebra and Databases, Ellis Horwood Series on Computers and their Applications, Wiley, Chichester, 1984.Google

37. [326]

Henschen, L.J., W.W. McCune and S.A. Naqvi, Compiling constraint checking programs from first order formulas, in: H. Gallaire , J.
Minker and J.M. Nicolas, eds.,Advances in Database Theory, Vol. 2, Plenum Press, New York, 1984, 145–169.Google Scholar

38. [327]

Jacobs, B.E., On database logic,Journal of ACM 29(2) (April 1982) 310–332.Google Scholar

39. [328]

Kellog, C., P. Klahr and L. Travis, Deductive planning and path finding for relational databases, in: H. Gallaire and J. Mink er,
eds.,Logic and Databases, Plenum Press, New York, 1978, 179–200.Google Scholar

40. [329]

Maier, D., A.O. Mendelzon and Y. Sagiv, Testing implications of data dependencies,ACM Transactions on Database Systems 4(4)
(1979) 455–469.Google Scholar

41. [330]

Maier, D., Y. Sagiv and M. Yannakakis, On the complexity of testing implications of functional and join dependencies,Journal of
ACM 28(4) (1981) 680–695.Google Scholar

42. [331]

Minker, J., An experimental relational data base system based on logic, in: H. Gallaire and J. Minker, eds.,Logic and Databases,
Plenum Press, New York, 1978, 107–147.Google Scholar

43. [332]

Naqvi, S.A. and L.J. Henschen, On compiling queries in recursive first order databases,Journal of ACM 31(1) (January 1984) 47–
85.Google Scholar
44. [333]

Reiter, R., Equality and domain closure in first order databases,Journal of ACM 27(2) (April 1980) 235–249.Google Scholar

45. [334]

Sacca, D., Closures of databases hypergraphs,Journal of ACM 32(4) (1985) 774–803.Google Scholar

46. [335]

Sagiv, Y., Quadratic algorithms for minimizing joins in restricted relational expressions,SIAM Journal on Computing 12(2) (1983)
316–328.Google Scholar

47. [336]

Sagiv, Y., An algorithm for inferring multivalued dependencies with an application to propositional logic,Journal of ACM 27(2) (1980)
250–262.Google Scholar

48. [337]

Sagiv, Y., C. Delobel, D.S. Parker and R. Fagin, An equivalence between relational database dependencies and a subclass of
propositional logic,Journal of ACM 28(3) (1981) 435–453.Google Scholar

49. [338]

Ullman, J.D.,Principles of Database Systems, 2nd ed., Computer Science Press, Potomac, MD, 1982.Google Scholar

50. [339]

Vardi, M.Y., The decision problem for database dependencies,Information Processing Letters 12(5) (1981) 251–254.Google Scholar

51. [340]

Warren, D.H.D., Efficient processing of interactive relational database queries expressed in logic,Proceedings 7th International
Conference on Very Large Data Bases, Cannes, France (1981) 272–281.Google Scholar

52. [341]

Fikes, R.E. and N.J. Nilsson, STRIPS: a new approach to the application of theorem proving to problem solving, Artificial
Intelligence 2 (1971) 189–208.Google Scholar

53. [342]

Fikes, R.E., P.E. Hart and N.J. Nilsson, Learning and executing generalized robot plans,Artificial Intelligence 3 (1972) 251–
288.Google Scholar

54. [343]

Hax, A.C. and J.J. Golovin, Hierarchical production planning systems, in: A.C. Hax, ed.,Studies in Operations Management, North-
Holland, Amsterdam, 1978.Google Scholar

55. [344]

Hayes-Roth, B. and F. Hayes-Roth, A cognitive model of planning,Cognitive Science 3 (1979) 275–310.Google Scholar

56. [345]

Hayes-Roth, B., A blackboard architecture for control,Artificial Intelligence 26 (1985) 251–321.Google Scholar

57. [346]

McDermott, D.V., Planning and action,Cognitive Science 2(2) (1978) 71–109.Google Scholar

58. [347]
Sacerdoti, E.D., Planning in a hierarchy of abstraction spaces,Proceedings of IJCAI 3(1973) 412–422.Google Scholar

59. [348]

Sacerdoti, E.D., The nonlinear nature of plans,Proceedings of IJCAI 4 (1975) 206–294.Google Scholar

60. [349]

Sacerdoti, E.D.,A Structure for Plans and Behavior, Elsevier, New York, 1977.Google Scholar

61. [350]

Stefik, M., Planning with constraints (MOLGEN: part 1),Artificial Intelligence 16 (1981) 111–140.Google Scholar

62. [351]

Stefik, M., Planning with constraints (MOLGEN: part 2),Artificial Intelligence 16 (1981) 141–170.Google Scholar

63. [352]

Wilensky, R.,Planning and Understanding: A Computational Approach to Human Reasoning, Addison-Wesley, MA, 1983.Google

64. [353]

Allen, J.F. and J.A. Koomen, Planning using a temporal world model,Proceedings of IJCAI 8 (1983) 742–747.Google Scholar

65. [354]

McDermott, D.V., A temporal logic for reasoning about processes and plans,Cognitive Science 6 (1982) 101–155.Google Scholar

66. [355]

Wilkins, D., Domain-independent planning: representation and plan generation,Artificial Intelligence 22 (1984) 269–301.Google

67. [356]

Corkill, D.D. and V.R. Lesser, The use of meta-level control for coordination in a distributed problem solving,Proceedings of IJCAI
8 (1983) 748–756.Google Scholar

68. [357]

Davis, R. and R.G. Smith, Negotiation as a metaphor for distributed problem-solving,Artificial Intelligence 20 (1983) 63–109.Google

69. [358]

Kornfeld, W.A., Combinatorially implosive algorithms,Communications of ACM 25(10) (October 1982) 734–738.Google Scholar

70. [359]

Lesser, V.R. and D.D. Corkill, Functionally-accurate, cooperative distributed systems,IEEE Transactions on Systems, Man and
Cybernetics SMC-11(1) (1981) 81–96.Google Scholar

71. [360]

Carbonell, J.G., R.S. Michalski and T.M. Mitchell, An overview of machine learning, in: R.S. Michalski, J.G. Carbonell and T.M.
Mitchell, eds.,Machine Learning, Tioga Press, Palo Alto, 1983, 3–23.Google Scholar

72. [361]

Farquhar, P.H., Application of utility theory in artificial intelligence research, in: Y. Sawaragi, ed. ,Multiple Criteria Decision-Making,
Springer-Verlag, New York, to appear.Google Scholar
73. [362]

Forsyth, R. and R. Dada,Machine Learning Applications in Expert Systems and Information Retrieval, Wiley, New York,
1986.Google Scholar

74. [363]

Langley, P., Learning to search: from weak methods to domain specific heuristics,Cognitive Science 9(2) (1985) 217–261.Google

75. [364]

Michalski, R.S., J.G. Carbonell and T.M. Mitchell, eds.,Machine Learning, Vol. 1, Tioga, Palo Alto, CA, 1983.Google Scholar

76. [365]

Michalski, R.S., J.G. Carbonell and T.M. Mitchell, eds.,Machine Learning, Vol. 2, Morgan Kaufmann, Los Altos, CA, 1986.Google

77. [366]

Mitchell, T.M., J.G. Carbonell and R.S. Michalski,Machine Learning: A Guide to Current Research, Academic Press, New York,
1986.Google Scholar

78. [367]

Breiman, L., J. Friedman, R. Olshun and C. Stone,Classification and Regression Trees, Wadsworth, Belmont, CA, 1984.Google

79. [368]

Fu, K.S., ed.,Applications of Pattern Recognition, CRC Press, Boca Ratón, FL, 1982.Google Scholar

80. [369]

Nandhakumar, N. and J.K. Aggarwal, The artificial intelligence approach to pattern recognition — a perspective and an
overview,Pattern Recognition 18(6) (1985) 383–389.Google Scholar

81. [370]

Quinlan, J.R., Inductive inference as a tool for the construction of high performance program, in: R. Michalski, J. Ca rbonell and T.M.
Mitchell, eds.,Machine Learning: An Artificial Intelligence Approach, Tioga Press, Palo Alto, CA, 1984.Google Scholar

82. [371]

Quinlan, J.R., Induction of decision trees,Machine Learning 1(1) (1986) 81–106.Google Scholar

83. [372]

Holland, J.H., K.J. Holyoak, R.E. Nisbett and P.R. Thagard,Induction: Processes of Inference, Learning and Discovery, M.I.T. Press,
Bradford, MA, 1986.Google Scholar

84. [373]

Johnson, M.K. and L. Hasher, Human learning and memory, in: M.R. Rosenzweig and L.W. Porter, eds.,Annual Review of
Psychology, Stanford, CA, 1987, 631–638.Google Scholar

85. [374]

Laird, J.E., P.S. Rosenbloom and A. Newell, Chunking in SOAR: the anatomy of a general learning mechanism,Machine Learning 1(1)
(1986) 11–46.Google Scholar

86. [375]

⋆ Lenat, D., EURISKO: A program that learns new heuristics and domain concepts. The nature of heuristics III: program design and
results,Artificial Intelligence 21(1–2) (1983) 61–98.Google Scholar
87. [376]

Mitchell, T.M., Generalization as search,Artificial Intelligence 18(2) (1982) 203–226.Google Scholar

88. [377]

Ackley, D.H. and G.E. Hinton, A learning algorithm for Boltzmann machines,Cognitive Science 9 (1985) 147–169.Google Scholar

89. [378]

Feldman, J.A. and D.H. Ballard, Connectionist models and their properties,Cognitive Science 6 (1982) 205–254.Google Scholar

90. [379]

McClelland, J.L., D.E. Rumelhart and PDP Research Group, eds.,Parallel Distributed Processing: Explorations in the Microstructure
of Cognition, M.I.T. Press, Cambridge, MA, 1985.Google Scholar

91. [380]

Grefenstette, J., ed.,Proceedings of the International Conference on Genetic Algorithms and Their Applications, Pittsburgh, PA,
1985.Google Scholar

92. [381]

Holland, J.H.,Adaptation in Natural and Artificial Systems, University of Michigan Press, Michigan, 1975.Google Scholar

93. [382]

Smith, S.F., Adaptive learning systems, in: R. Forsyth, ed.,Expert Systems, Principles and Case Studies, Chapman and Hall, Ltd., New
York, 1984.Google Scholar

94. [383]

Buchanan, B.G., Research on expert systems,Machine Intelligence, Wiley, New York, 1982, 269–300.Google Scholar

95. [384]

Duda, R.O. and E.H. Shortliffe, Expert systems research,Science 220 (1983) 262–268.Google Scholar

96. [385]

Hayes-Roth, F., D.A. Waterman and D. Lenat, eds.,Building Expert Systems, Addison-Wesley, Reading, MA, 1983.Google Scholar

97. [386]

Klahr, P., and D.A. Waterman, eds.,Expert Systems, Addison-Wesley, Reading, MA, 1986.Google Scholar

98. [387]

Michie, D., ed.,Expert Systems in the Micro-Electronic Age, Edinburgh University Press, Edinburgh, 1980.Google Scholar

99. [388]

Reitman, W., ed.,Artificial Intelligence Applications for Business, Ablex, Norwood, NJ, 1984.Google Scholar

100. [389]

Stefik, M., J. Aikins, R. Balzer, J. Benoit, L. Birnbaum, F. Hayes-Roth and E.H. Sacerdoti, The organization of expert systems: a
tutorial,Artificial Intelligence 18 (1982) 135–173.Google Scholar

101. [390]

Waterman, D.A., ed.,A Guide to Expert Systems, Addison-Wesley, Reading, MA, 1986.Google Scholar

102. [391]
Weiss, S.M. and C.A. Kulikowski,A Practical Guide to Designing Expert Systems, Rowman and Allanheld, Totowa, NJ, 1984.Google

103. [392]

Winston, P.H. and K.A. Prendergast,The AI Business, M.I.T. Press, Bradford, MA, 1984.Google Scholar

104. [393]

Fischer, E.L., An AI-Based methodology for factory design,AI Magazine (Fall 1986) 72–85.Google Scholar

105. [394]

Fox, B.R. and K.G. Kempf, Opportunistic scheduling for robotic assembly,Proceedings IEEE International Conference on Robotics
and Automation, 1985, 880–889.Google Scholar

106. [395]

Fox, M.S. and S.F. Smith, ISIS: a knowledge-based system for factory scheduling,Expert Systems 1 (1984) 25–49.Google Scholar

107. [396]

Gilmore, J.F. and A.C. Semeco, Terrain navigation through knowledge-based route planning,Proceedings of IJCAI 9 (1985) 1086–
1088.Google Scholar

108. [397]

Goldstein, I.P. and R.B. Roberts, NUDGE — a knowledge-based scheduling program,Proceedings of IJCAI 5 (1977) 257–263.Google

109. [398]

Kastner, J., C. Apte, J. Griesmier, S.J. Hong, M. Karnaugh, E. Mays and Y. Tozawa, A knowledge-based consultant for financial
marketing,AI Magazine 7(5) (Winter 1986) 71–79.Google Scholar

110. [399]

Lynch, F., C. Marshall, D.E. O'Connor and M. Kiskiel, AI in manufacturing at Digital,AI Magazine 7(5) (Winter 1986) 53–57.Google

111. [400]

⋆ McDermott, D., A temporal logic for reasoning about processes and plans,Cognitive Science 6 (1982) 101–155.Google Scholar

112. [401]

Miller, D., R.J. Firby and T. Dean, Deadlines, travel time and robot problem solving,Proceedings of IJCAI 9 (1985) 1052–1054.Google

113. [402]

Murphy, F.H. and E.A. Stohr, An intelligent system for formulating linear programs,Decision Support Systems 2(1) (March 1986) 39–
47.Google Scholar

114. [403]

O'Connor, D.E., Using expert systems to manage change and complexity in manufacturing, in: W. Reitman, ed., Artificial Intelligence
Applications for Business, Ablex, Norwood, NJ, 1984.Google Scholar

115. [404]

Orciuch, E. and J. Frost, ISA: an intelligent scheduling assistant,IEEE Computer Society, Proceedings 1st Conference on AI
Applications, 1984.Google Scholar

116. [405]
Ow, P.S. and S.F. Smith, Viewing scheduling as an opportunistic problem-solving process,this volume.Google Scholar

117. [406]

⋆ Paul. R.J. and G.I. Doukidis, Further developments in the use of artificial intelligence techniques which formulate simulation
problems,Journal of the Operational Research Society 37(8) (1986) 787–810.Google Scholar

118. [407]

Sathi, A., T.E. Morton and S.F. Roth, Callisto: an intelligent project management system,AI Magazine 7(5) (Winter 1986) 34–
52.Google Scholar

119. [408]

Scarl, E., Applications of a non-rule knowledge-based system to NASA scheduling and monitoring problems,IECEC Proceedings San
Francisco (August 1984).Google Scholar

120. [409]

Smith, S.F., M.S. Fox and P.S. Ow, Constructing and maintaining detailed production plans: investigations into the development of
knowledge-based factory scheduling systems,AI Magazine (Fall 1986) 45–61.Google Scholar

121. [410]

Sugie, M., D. Menzilcioglu and H.T. Kung, CAR guide — onboard computer for automobile route guidance,Proceedings National
Computer Conference, 1984, 695–706.Google Scholar

122. [411]

Tonge, F.M., Summary of a heuristic line balancing procedure, in: E.A. Feigenbaum and J. Feldman, eds., Computers and Thought,
McGraw-Hill, New York, 1983.Google Scholar

123. [412]

Buchanan, B.G. and E.A. Feigenbaum, DENDRAL and META-DENDRAL,Artificial Intelligence 11 (1978) 5–24.Google Scholar

124. [413]

Engelmore, R. and A. Terry, Structure and function of the CRYSALIS system,Proceedings of IJCAI 6 (1979) 250–256.Google Scholar

125. [414]

Erman, L.D., F. Hayes-Roth, V.R. Lesser and D.R. Reddy, The HEARSAY-II speech understanding system: integrating knowledge to
resolve uncertainty,Computing Surveys12(2) (June 1980) 213–253.Google Scholar

126. [415]

McCracken, D., Representation and efficiency in a production system for speech understanding,Proceedings of IJCAI 6 (1979) 556–
561.Google Scholar

127. [416]

Campbell, A., V. Hollister, R.O. Duda and P.E. Hart, Recognition of a hidden mineral deposit by an artificial intelligence
program,Science 217 (1982) 927–929.Google Scholar

128. [417]

Davis, R., H. Austin, I. Carlbom, B. Frawley, P. Pruchnik, R. Sneiderman and J. Gilreath, The DIPMETER ADVISOR
system,Proceedings of IJCAI 8 (1983) 122–129.Google Scholar

129. [418]

Descotte, Y. and J.C. Latombe, GARI: a problem solver that plans how to machine mechanical parts,Proceedings of IJCAI 7 (1981)
776–772.Google Scholar
130. [419]

Gallanti, M., G. Guida, L. Spampinato and A. Stefanini, Representing procedural knowledge in expert systems: an application to
process control,Proceedings of IJCAI 9(1985) 345–352.Google Scholar

131. [420]

Geneseretti, M.R., The use of design descriptions in automated diagnosis,Artificial Intelligence 24 (1984) 411–436.Google Scholar

132. [421]

Konolige, K., An inference net compiler for the PROSPECTOR rule-based consultation system,Proceedings of IJCAI 6 (1979) 487–
489.Google Scholar

133. [422]

Weiss, S., C. Kulikowski, C. Apte, M. Uschold, J. Patchette, R. Brigham and B. Spitzer, Building expert systems for controlli ng complex
programs,Proceedings of NCAI 2, Pittsburgh, PA, 1982, 322–326.Google Scholar

134. [423]

Buchanan, B.G. and E.H. Shortliffe, (eds.),Rule-Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic
Programming Project, Addison-Wesley, Reading, MA, 1985.Google Scholar

135. [424]

Clancey, W.J. and R. Letsinger, NEOMYCIN: reconfiguring a rule-based expert system for application to teaching,Proceedings of
IJCAI 7 (1981) 829–836.Google Scholar

136. [425]

Frienland, P., Knowledge-based experiment design in molecular genetics,Proceedings of IJCAI 6 (1979) 285–288.Google Scholar

137. [426]

Shortliffe, E.H., A.C. Scott, M.B. Bischoff, A.B. Campbell, W. Van Mellerd and C.D. Jacobs, ONCOCIN: an expert system for onc ology
protocol management,Proceedings of IJCAI 7 (1981) 876–881.Google Scholar

138. [427]

Balzer, R., L. Erman, P. London and C. Williams, HEARSAY-III: a domain-independent framework for expert systems,Proceedings of
NCAI 1 (August 1980) 108–110.Google Scholar

139. [428]

Feigenbaum, E.A., B.G. Buchanan and J. Lederberg, On generality and problem solving: a case study using the DENDRAL program,
in: B. Meltzer and D. Michie, eds.,Machine Intelligence 6, Elsevier, New York, 1971.Google Scholar

140. [429]

Neches, R., W.R. Swartout and J. Moore, Explainable (and maintainable) expert systems,Proceedings of IJCAI 9 (1985) 382–
389.Google Scholar

141. [430]

Nii, H.P., J.J. Anton, E.A. Feigenbaum and A.J. Rockmore, Signal-to-symbol transformation: HASP/SIAP case study,AI
Magazine (1982) 23–35.Google Scholar

142. [431]

Ow, P.S. and S.F. Smith, Two design principles for knowledge-based systems,Decision Sciences (July 1987).Google Scholar

143. [432]

Quinlan, J., Knowledge-based system for locating missing high cards in bridge,Proceedings of IJCAI 6 (1979) 705–707.Google Scholar
144. [433]

Swartout, W.R., Explaining and justifying expert consulting programs,Proceedings of IJCAI 7 (1981) 815–823.Google Scholar

145. [434]

Bachant, J. and J. McDermott, R1 revisited: four years in the trenches,AI Magazine (Fall 1984) 21–32.Google Scholar

146. [435]

McDermott, J., R1's formative years,AI Magazine 2(2) (Summer 1981) 21–29.Google Scholar

147. [436]

McDermott, J., R1: a rule-based configurer of computer systems,Artificial Intelligence19(1) (September 1982) 39–88.Google Scholar

148. [437]

McDermott, J., Domain knowledge and the design process,Design Studies 3(1) (January 1982).Google Scholar

149. [438]

Polit, S., R1 and beyond: technology transfer at DEC,AI Magazine (Winter 1985) 76–78.Google Scholar

150. [439]

Rosenbloom, P.S., J.E. Laird, J. McDermott, A. Newell and E. Orciuch, R1-SOAR: an experiment in knowledge-intensive programming
in a problem-solving architecture,IEEE Transactions on Pattern Analysis and Machine Intelligence 7(5) (1985) 561–569.Google

151. [440]

Scown, S.J.,The Artificial Intelligence Experience: An Introduction, Digital Equipment Corporation, 1985.Google Scholar

152. [441]

Van de Brug, A., J. Bachant and J. McDermott, The taming of R1,IEEE Expert (Fall 1986) 33–39.Google Scholar

153. [442]

Bonissone, P.P., Plausible reasoning: coping with uncertainty in expert systems, in: S.C. Shapiro, ed.,Encyclopedia of Artificial
Intelligence, Wiley, New York, 1986.Google Scholar

154. [443]

Cheeseman, P., In defense of probability,Proceedings of 9th International Joint Conference on AI, Los Angeles, CA, 1985.Google

155. [444]

Henrion, M., Uncertainty in artificial intelligence: is probability epistemologically and heuristically adequate?Expert Systems and
Expert Judgement, Proceedings of the NATO Advanced Research Workshop, Porto, Portugal, August 1986.Google Scholar

156. [445]

Kanal, L.N. and J. Lemmer, eds.,Uncertainty in AI, North-Holland, Amsterdam, 1986.Google Scholar

157. [446]

Spiegelhalter, D.J., A statistical view of uncertainty in expert systems, in: W. Gale, ed.,AI and Statistics, Addison-Wesley, Reading,
MA, 1986, 17–55.Google Scholar

158. [447]
Duda, R.O., P.E. Hart and N.J. Nilsson, Subjective Bayesian methods for rule-based inference systems,AFIPS Conference Proceedings,
New York, June 1976, 1075–1082.Google Scholar

159. [448]

Jeroslow, R.G., On monotone chaining procedures of the CF type, Technical Report, College of Management, Georgia Institute of
Technology, 1985.Google Scholar

160. [449]

Negoita, C.V.,Expert Systems and Fuzzy Systems, Benjamin/Cummings Publishing Co., Menlo Park, CA, 1985.Google Scholar

161. [450]

Pearl, J., Fusion, propagation, and structuring belief networks,Artificial Intelligence 29(3) (September 1986) 241–288.Google Scholar

162. [451]

Shafer, G.,A Mathematical Theory of Evidence, Princeton University Press. Princeton, 1976.Google Scholar

163. [452]

Shachter, R.D., Evaluating influence diagrams,Operations Research (to appear in 1987).Google Scholar

164. [453]

Shortliffe, E.H. and B.G. Buchanan, A model of inexact reasoning in medicine, in: B.G. Buchanan and E.H. Shortliffe, eds., Rule-Based
Expert Systems, Addison-Wesley, Reading, MA, 1984, 233–262.Google Scholar

165. [454]

Zadeh, L.A., Making computers think like people,IEEE Spectrum (August 1984) 26–32.Google Scholar

166. [455]

Tong, R.M. and D.G. Shapiro, Experimental investigations of uncertainty in a rule-based system for information
retrieval,International Journal of Man-Machine Studies 17 (March 1985) 265–282.Google Scholar

167. [456]

Wise, B., An experimental comparison of uncertain inference systems, Ph.D. thesis, Department of Engineering and Public Polic y,
Carnegie-Mellon University, 1986.Google Scholar

168. [457]

Yadrick, R.M., D.S. Vaughan, B.M. Perrin, P.D. Kempf and K.G. Kempf, Evaluation of uncertain interference models I:
PROSPECTOR,2nd Workshop on Uncertainty AAAI, 1986.