Expert systems in

medicine
Concept of ES
Expert System - an intelligent
computer program that uses
knowledge and inference
procedures to solve enough
challenges and requires for its
solution a significant amount of
expertise human
n Expert System - a computer system that
emulates the expert's ability to make
decisions.
n Expert System is obliged to act in all respects
as expert persons.
n The concept of emulation like a simulation,
but simulation system is required to act like a
real object only in certain respects.
History
n The term entered E. Feigenbaum and his
interpretation is "the introduction of the
principles and tools of artificial intelligence
applied to the solution of difficult problems
that require expert knowledge.
n Field studies devoted formalization
methods of knowledge representation and
construction of expert systems (ES), is called
"knowledge engineering".
EC intended for so-called non-
formalized tasks, such as:
ninterpretation,
nprediction,
ndiagnostics,
nplanning,
nconstruction,
ncontrol,
ndebugging,
ncoaching,
ncontrol
The principles of
construction of ES
The user transmits an expert
system facts or other
information, and receives as a
result of expert advice or
expertise
The principles of
construction of ES
n In its structure, the expert system is
divided into two main components - the
knowledge and inference machine
base.
n The knowledge base contains the
knowledge on which the inference
machine generates a conclusion.
Scheme of ES

Expert system

Data
data base

USER

Expert knowledge
Block
for
output
 Specialization of ES

n A special area of concern, such as

medicine, finance, science and
technology, which could very well
solve the problem of a certain
expert
Elements for organization ES
n solver (interpreter);
n working memory, also called the database;
n Knowledge Base (KB);
n components of knowledge acquisition;
n explanatory component;
n interactive component
n Database (working memory) is intended for
storage of raw and intermediate data are
being solved at the moment
n Knowledge Base (KB) in the ES is designed
for long-term storage of data describing the
treated area and the rules that describe the
transformation of suitable data in the area
n Solver, using the raw data from the working
memory and the knowledge of the base,
forming a sequence of rules to the solution
of the problem.
Performer for ES
n The expert: determines the knowledge (data and
rules), characterizing the problem area, ensures their
completeness and correctness
n The knowledge engineer helps identify and structure
the knowledge necessary for ES; It provides a choice
of ES, identifies and programs the standard features
to be used in the rules introduced by the expert.
n The programmer develops program contained within
all the main components of ES, and carries out its
interface with the environment in which it will be
used.
n User - specialist subject area, which is intended for
system
ES can operate in two modes:
n Input mode of knowledge - in this mode with the help
of an expert engineer knowledge through knowledge
base editor enters information known to him about
the subject area in the ES knowledge base.
n Consultation mode - user conducts a dialogue with
the EC, telling her information about the current task
and receiving EC recommendations. For example,
based on information about the physical condition of
the patient makes the diagnosis of ES in the form of a
list of diseases, the most likely under these
symptoms.
The type ES in medicine
n By way of solutions:
n Probability
n Expert
n According to the type of support solutions:
n systems that improve the diagnosis,
n the system offers the best treatment strategy

n By type of intervention:
n Passive
n Semi-active
n Active
Artificial Intelligence in Medicine
n Artificial intelligence is technology-science
and the creation of intelligent machines

n AI is associated with a similar task of using

computers to understand human intelligence,
but not necessarily limited to biologically
plausible methods
 History

n The term introduced in 1956 at Stanford

University
n Since 1940, the birth of AI belongs to N. Wiener.
n 1954 Moscow State University Professor
Lyapunov A.A. work in the field of "Machines and
thinking" in the field psychology and
mathematics.
n In the late '70s of the 20th century. Modeling
knowledge bases specific professional experts
n 80-90yy Development of knowledge
representation languages, expert systems
Two directions for creating AI
n (Top-Down AI) is n Bottom-Up AI) - the
creation of expert study of neural
systems, and networks, and
knowledge without evolutionary computed,
inference systems, modeling intelligent
imitating the high-level behavior based on
n Its mental processes: biological elements, as
thinking, reasoning, well as the creation of
speech, emotions appropriate computer
systems, such
neurocomputer or bio-
computer.
Areas of use
n medical diagnostics
n control and monitoring
n diagnosis of faults in mechanical
and electrical devices
n training
The classes of problems solved with
the use of AI
n real-time monitoring
n control systems
n fault detection system
n diagnostics, scheduling
n planning, optimization
n system-advisers
n system design.
Technology of AI
n Data mining (Data Mining)
n Neural networks
n reasoning systems
n Systems for detection logic patterns in the
data
n genetic algorithms
Medical diagnostic (MYCIN)
n Diagnostic systems are used
to establish the connection
between impaired activity of
the organism and their
possible causes.
n it is intended for the
diagnosis and monitoring of
the condition of the patient
with meningitis and bacterial
infections.
n Currently, this system makes
the diagnosis to a specialist
level.
n She has extensive
knowledge base, so that can
be used in other areas of
medicine.
Desktop of AI
Biological modeling of
artificial intelligence
n artificial systems are not required to be
repeated in the structure and function of
the structure and processes occurring in
it, inherent in biological systems.
Examples
n Neural networks are used to solve a complex
issues such as the recognition of geometric
shapes or clustering objects.
n The genetic algorithms is based on the idea
that an algorithm can be more efficient if the
best characteristics borrow other algorithms
("parents").
n A relatively new algorithms, which seeks to
create a stand-alone program - an agent that
interacts with the external environment
Existing expert systems
n CLIPS — famous popular application for creating
ES (public domain)
n OpenCyc — power and dynamic ES with
global ontological model and support
independent contexts
n WolframAlpha —search engine, intelligent
"computational knowledge engine"
 Modern technologies
which used in Expert system

Artificial neural
networks
n Nonlinear system that
allows you to classify
the data much better
than the commonly
used linear methods.
n Attached to the medical
diagnostics enable
significantly improve the
specificity of the
method, without
reducing its sensitivity
Conclusion
n ES application given big enables the
doctor did not take the time to address
some routine and mentally challenging
tasks
n Artificial intelligence is a step in the
future, new technologies as diagnostic
and treatment