Artificial Intelligence Diagnosis: Fundamentals and Applications

By Fouad Sabry

What Is Artificial Intelligence Diagnosis

Diagnosis is an area of artificial intelligence that focuses on the development of algorithms and methods that are capable of determining whether or not the behavior of a system is appropriate. In the event that the system is not operating as it should, the algorithm should be able to identify, with as much precision as is feasible, the component of the system is malfunctioning as well as the nature of the problem that it is experiencing. The computation is founded on observations, which supply information on the behavior that is now taking place.

How You Will Benefit

(I) Insights, and validations about the following topics:

Chapter 1: Diagnosis (artificial intelligence)

Chapter 2: Inductive logic programming

Chapter 3: Machine learning

Chapter 4: Intelligent agent

Chapter 5: Artificial intelligence in healthcare

Chapter 6: Symbolic artificial intelligence

Chapter 7: Internist-I

Chapter 8: Model-based reasoning

Chapter 9: Partially observable Markov decision process

Chapter 10: Fault detection and isolation

(II) Answering the public top questions about artificial intelligence diagnosis.

(III) Real world examples for the usage of artificial intelligence diagnosis in many fields.

(IV) 17 appendices to explain, briefly, 266 emerging technologies in each industry to have 360-degree full understanding of artificial intelligence diagnosis' technologies.

Who This Book Is For

Professionals, undergraduate and graduate students, enthusiasts, hobbyists, and those who want to go beyond basic knowledge or information for any kind of artificial intelligence diagnosis.

• Language: English
• Publisher: One Billion Knowledgeable
• Release date: Jul 4, 2023

Book preview

Chapter 1: Diagnosis (artificial intelligence)

Diagnosis is a subfield of artificial intelligence that focuses on creating algorithms and techniques to test the validity of a system's behavior. If the system is malfunctioning, the algorithm should be able to pinpoint the exact location of the problem and identify the nature of the error with high precision. The computation relies on observations, which reveal details about the behavior in the present.

Diagnosis can also mean both the process of determining whether or not a system is broken and the determination itself. The medical definition of this word is the act of determining the nature of an illness from a patient's presentation of symptoms.

The work of an auto repair shop's mechanic is a good illustration of the diagnostic process in action. The mechanic will initially make use of his familiarity with the vehicle and his observations to see if anything out of the ordinary has occurred. The mechanic's role in the vehicle diagnosis cannot be overstated; should he discover abnormal behavior, he will try to refine his diagnosis using new observations and possibly testing the system until he discovers the faulty component.

The expert diagnosis, also known as a diagnosis by expert system, is grounded in the clinician's familiarity with the software. Based on this information, a mapping is constructed that accurately links symptoms with their diagnoses.

Expertise in this area is available:

by a real live person. In this situation, human expertise needs to be codified in a computer language.

By showing how the system acts in specific cases. The examples here need to be labeled as either correct or incorrect (and, in the latter case, by the type of fault). The data is then generalized using machine learning techniques.

Major limitations of these approaches include:

The challenge of learning the material. Long-term use of the system is usually required before such knowledge becomes available (or similar systems). As a result, these techniques can't be used in systems where safety or mission success is paramount (such as a nuclear power plant, or a robot operating in space). In addition, there is no way to know for sure that the acquired expert knowledge is exhaustive. It is impossible to make a diagnosis in the event of a sudden onset of symptoms or the manifestation of a new symptom.

Learning's inherent complexity. Expert system development occurs off-line, which can eat up a lot of processing time and data storage space.

The total storage capacity of the expert system. An excessive amount of space may be needed as the expert system attempts to map any observation to a diagnosis.

The fragility of the system. The entire process of building the expert system must be redone if any changes are made, no matter how minor.

Expert systems can also be constructed in a slightly different way, by starting with a model of the system rather than with a specific body of knowledge. In order to diagnose discrete-event systems, for instance, a diagnoser must be computed. Although it has some similarities to model-based methods, this strategy also shares some strengths and weaknesses with expert system approaches.

Diagnosis based on a model is an application of abductive reasoning. In a nutshell, it operates as follows::

Principle of the model-based diagnosis

The system's behavior is captured in a model that we've constructed (or artefact). The model is an approximation of the system's actual behavior and as such may have flaws. In particular, the incorrect behavior is usually obscure, and the incorrect model may therefore be absent. The diagnosis system takes incoming data, simulates the system according to the model, and then compares the simulated results to the actual results.

The modelling can be simplified by the following rules (where Ab\, is the Abnormal predicate):

\neg Ab(S)\Rightarrow Int1\wedge Obs1

Ab(S)\Rightarrow Int2\wedge Obs2 (fault model)

These formulas have the following meaning: Providing normal system behavior (i.e.

if it is typical), then the internal (unobservable) behaviour will be Int1\, and the observable behaviour Obs1\, .

Otherwise, the internal behaviour will be Int2\, and the observable behaviour Obs2\, .

Given the observations Obs\, , the problem is to determine whether the system behaviour is normal or not ( \neg Ab(S)\, or Ab(S)\, ).

Abductive reasoning is demonstrated here.

A system is said to be diagnosable if, regardless of the system's behavior, a clear diagnosis can be made.

Since there are two competing goals—reducing the number of sensors to cut costs and increasing the number of sensors to increase the likelihood of detecting a faulty behavior—the problem of diagnosability is crucial when designing a system.

There are multiple algorithms available for solving such issues. Is the system diagnosable? is answered by one set of algorithms, while another set searches for sensor combinations that make the system diagnosable and, optionally, meet criteria like cost optimization.

In most cases, the diagnosability of a system can be calculated using the system model. Such a model does not need to be constructed from scratch and can be used immediately in model-based diagnosis applications.

{End Chapter 1}

Chapter 4: Inductive logic programming

Inductive logic programming (ILP) is a branch of symbolic AI that standardizes on logic programming to represent data like data sets, knowledge bases, and hypotheses. An ILP system, given a knowledge encoding and a set of examples in the form of a logical database of facts, will generate a hypothesized logic program that includes all the positive and none of the negative examples.

Schema: positive examples + negative examples + background knowledge ⇒ hypothesis.

Bioinformatics and NLP are two fields that benefit greatly from inductive logic programming. In a logical context, Gordon Plotkin and Ehud Shapiro established the first theoretical groundwork for inductive machine learning. In the PROGOL system, Muggleton was the first to implement Inverse entailment. In this context, induction refers more to philosophical than mathematical induction (the latter being the process of proving a property