Towards reasoning visualization in expert systems

SASHA is a model-driven application shell which generates expert systems for fault isolation and repair in complex electronic systems. SASHA incorporates an expert system shell and a knowledge acquisition tool into a unique mechanism that elicits data directly from the domain expert and outputs a rule-based diagnostic expert system.SASHA employs causal-model and rule-based representations. It uses a causal model to identify the structure, behavior and relation of a domain object. It uses rules to direct the diagnostic problem solver. SASHA's initial knowledge base consists of a domain independent generalized model of the system under analysis. This underlying model drives knowledge acquisition and the consequent generation of an expert system. SASHA introduces a new concept of Pattern Rules, which are a generalized skeleton of the diagnostic problem solver. During diagnosis, when applied to the domain specific Structural Knowledge base, Pattern Rules dynamically instantiate the actual domain rules. Based on the Pattern Rules notion, three types of rule base organization are proposed and analyzed: Fully Patternized Organization (FPO), Partly Patternized Organization (PPO) and Customized Domain-specific Organization (CDO).

Development of a prototype expert system for identification and control of insect pests

Introduction. Expert witness reports, prepared with the aim of quantifying fault rates among parties, play an important role in a court's final decision. However, conflicting fault rates assigned by different expert witness boards lead to iterative objections raised by the related parties. This unfavorable situation mainly originates due to the subjectivity of expert judgments and unavailability of objective information about the causes of accidents. As a solution to this shortcoming, an expert system based on a rule-based system was developed for the quantification of fault rates in construction fall accidents. The aim of developing DsSafe is decreasing the subjectivity inherent in expert witness reports. Methodology. Eighty-four inspection reports prepared by the official and authorized inspectors were examined and root causes of construction fall accidents in Turkey were identified. Using this information, an evaluation form was designed and submitted to the experts. Experts were asked to evaluate the importance level of the factors that govern fall accidents and determine the fault rates under different scenarios. Based on expert judgments, a rule-based expert system was developed. The accuracy and reliability of DsSafe were tested with real data as obtained from finalized court cases. Result. DsSafe gives satisfactory results.

Rule-based systems are relatively easier to explain to end users because rules can be expressed as if/else conditions. Indeed, at times there are multiple if/else conditions. Expert systems are deterministic in nature and very accurate, because they are driven by well-established rules and conditions. The need for explainability arises when multiple rules either conflict or are combined in such a way that it is difficult to interpret. In the field of artificial intelligence, an expert system is a program designed to simulate a real decision making process. Expert systems are designed to solve complex problems by simple reasoning and are represented as if/else/then conditions. Expert systems are designed to act like an inference engine. The role of an inference engine is to make decisions. Sometimes expert systems are simple, like a decision tree, and sometimes expert systems are very complex. Complex expert systems require a knowledge base. Sometimes these knowledge bases are an ontology, which is a form of a body of knowledge, through which a particular rule can be framed and inferencing can be done.

Ameliorating the Pregnant Man problem: a verification tool for personal computer based expert systems

Cost and gross profit calculations of agricultural products are a comprehensive and difficult process. Many different parameters, such as soil cultivation, pruning, spraying, irrigation, storage, drying, harvesting, transportation, and marketing, need to be evaluated together. Currently, cost calculations of agricultural products are made with classical programs and simple mathematical methods. For this reason, the results vary depending on the person and organization making the evaluation. This creates some limitations, making cost planning, which is a dynamic process, unreliable. In the study, a rule-based expert system was developed for the cost and gross profit calculation of apricots. The expert system developed is called Apricot Information System (APRIS). The developed expert system aims to provide more accurate solutions by adding the precision and speed of the computer to the calculation process. The main contribution of this study is the development of a complete decision support set covering all processes from tillage to marketing and the design of rule-based user-system interaction. The designed system found the apricot variable cost calculation to be 1126.36 TL/da for 2020. This value largely overlaps with the values obtained from the field in the previous year. The developed expert system will be a platform that can be used in cost calculations of many agricultural products.

Computer-Aided Knowledge-Based Monitoring and Diagnostic System for Emulsion Polymerization

This paper presents a planning methodology using an application of a mathematical and a rule-based expert system (ES) to expand the transmission network in a deregulated environment. In this methodology, the ES suggests a realistic set of generation additions with proper economic signals to the participants, before proceeding with the transmission expansion. A feasible list of transmission alternatives is then assumed to accommodate the proposals for generation. A mathematical method is performed based on marginal cost allocation to optimize the location for the new generation and its transmission expansion scheme simultaneously for each alternative. The optimum alternative that minimizes the overall system's cost function and satisfies the future demand is obtained. The ES interacts with the power system planning tools to produce the optimum expansion plan. A practical application is given to demonstrate the effectiveness of the developed prototype system.

Background Large language models (LLMs) such as GPT-4 can interpret free text, but unreliable answers, opaque reasoning, and privacy risks limit their use in healthcare. In contrast, rule-based artificial intelligence (AI) provides transparent and reproducible results but struggles with free text. We aimed to combine the strengths of both approaches to test whether such a hybrid system can autonomously and reliably extract clinical data from diagnostic imaging reports. Methods We developed a neuro-symbolic AI that connects GPT-4 with a rule-based expert system through a semantic integration platform. GPT-4 extracted candidate facts from free-text reports, while the expert system verified them against medical rules, producing traceable, deterministic labels. We evaluated the system on 206 consecutive prostate cancer PET/CT scan reports, requiring extraction of 26 clinical parameters per report, generating 5356 data points, and answering three study questions: study inclusion, recurrent cancer identification, and prostate-specific antigen (PSA) level retrieval. Outputs were compared against physician-derived references, and discrepancies were reviewed by a blinded adjudicator. Results Here we show that neuro-symbolic AI outperforms GPT-4 alone and matches physicians in structuring and analysing reports. GPT-4 alone achieves F1 scores of 0.63 for study inclusion and 0.95 for recurrence detection, with 96.6% correct PSA values. Physicians reach F1 scores of 1.00 and 0.99, with 98.1% PSA accuracy. The neuro-symbolic AI scores twice 1.00 with 100% PSA accuracy and delivers always an auditable chain of reasoning. It intercepts two intentionally introduced reports with residual identifiers, preventing unintended transfer of sensitive data. Conclusions Unlike standalone LLMs, neuro-symbolic AI can safely automate data extraction for clinical research and may provide a path toward trustworthy AI in healthcare practice.Medical doctors often write reports as free text, which is hard to reuse for research or care. A large language model is software that reads and writes text by imitating large networks of brain cells. This type of artificial intelligence can extract and organize important information from medical reports. But its reasoning is opaque, answers can be wrong, and it raises privacy concerns. Rule-based artificial intelligence is transparent, responds correctly, and is privacy-protecting but struggles with free text. We combined both artificial intelligence types, so each offsets the other’s weaknesses. We tested the system on 206 prostate cancer imaging reports, where it extracted information correctly, showed how it reached its answers, and protected sensitive data. Pairing large language models with rule-based systems could make artificial intelligence safer, more trustworthy, and more useful in healthcare.

Artificial intelligence (AI) techniques using sophisticated computer software are increasingly being used in the analysis of biological systems and for the design of biological models. The term artificial intelligence refers to machine-based processes that are functionally equivalent to human cognitive tasks for analyzing data in the domain for which they were designed. They are considered intelligent because of human-like reasoning processes. The expert system and the artificial neural network are two examples of such techniques (1). We are in the process of adapting these new computer information processing techniques into a knowledge-based modelling system which will be used to selectively search, analyze, retrieve and display thermoregulatory data in response to operator requests. We have developed a prototype system in which a rule-based expert system integrates its knowledge base with our accumulated historical database of experimentally obtained information on thermoregulation to provide new retrieval and analysis capabilities which will be useful for both research and teaching. This enables you to address otherwise unmanageable problems. Heuristics are often used. These types of systems are able to work with uncertainty, reason about what is unknown, and handle multivalued information with multiple conclusions. They are very effective at handling the problems that they were designed to solve, and are excellent for situations in which many people require access to a body of knowledge. Expert systems or neural networks when used alone are not optimal. The deficits in each can be filled by designing an integrated hybrid system in which neural networks are embedded within the rule based expert system and so both synergistically fill the gaps in each others technology.

The problem of verifying external adequacy in expert systems is addressed in this paper. External adequacy is defined in terms of a formal specification of question relevance, that expresses natural data dependencies among the external facts used in the ontology of the domain of expertise. A method is presented for verifying the expert system, at any stage of development, against the formal specification. The method is based on transforming the knowledge base plus inference engine into a partially ordered set of support graphs. The complex problem of verifying the external adequacy of expert systems is then translated into a simpler path-finding problem in support graphs.

Expert systems are built to solve a wide range of problems in domains such as medicine, science, engineering, manufacturing, education, and many other fields. A Rule-Based Expert System has the following components: the knowledge base, the inference engine, and a user interface that communicates with users. A Decision Tree is a hierarchical data structure consisting of nodes, which store information or knowledge, and edges, which connect the nodes. In this paper, we discuss that the inference or reasoning process of Expert System can be easier to identify by a decision tree structure through a forward reasoning inference or a backward reasoning inference. Forward reasoning inference is data-driven; its reasons for facts to find out what solutions (or conclusions) follow from the facts. However, backward reasoning inference is goal-driven, which needs to find facts that support the hypothesis. A small animal classification example will be discussed. For learning expert system programming effectively, we introduce two software tools and discuss their advantages. Firstly, CLIPS (C Language Integrated Production System) is discussed. CLIPS was designed using C programming language at the NASA/Johnson Space Center with the specific purpose of providing high portability, low cost, and easy integration with external systems. Secondly, an interactive logic programming tool, Visirule is discussed; VisiRule was offered by LPA, a small, private UK company specializing in practical Artificial Intelligence. Finally, we propose a Rubric for Outcome Assessments for learning expert system programming with decision Trees for anyone who is interested in practical expert system programming.

In this research, a prototype rule-based expert system for travel demand management (TDM) in selecting suitable policy was designed and developed. This expert system suggests a process for travel demand management policy implementation, and offers guidance and advice on the selection of effective and appropriate policies. The system is supported with an object-oriented modeling in the knowledge abstraction and representation, and the interactions with the system are performed through a friendly graphical user interface (GUI). The prototype of the system was developed in shell expert system platform (KAPPA-PC). The outcomes obtained by the current prototype show that the used rule-based expert system architecture resulted in successful computerized intelligent decision support software that is easy to maintain, modify, and extend.

In sports and rehabilitation processes where isotonic movements such as bodybuilding are performed, it is vital for individuals to be able to correct the wrong movements instantly by monitoring the trainings simultaneously, and to be able to train healthily and away from the risks of injury. For this purpose, we designed a new real-time athlete support system using Kinect V2 and Expert System. Lateral raise (LR) and dumbbell shoulder press (DSP) movements were selected as examples to be modeled in the system. Kinect V2 was used to obtain angle and distance changes in the shoulder, elbow, wrist, hip, knee, and ankle during movements in these movement models designed. For the rule base of Expert System developed according to these models, a 28-state rule table was designed, and 12 main rules were determined that could be used for both actions. In the sample trainings, it was observed that the decisions made by the system had 89% accuracy in DSP training and 82% accuracy in LR training. In addition, the developed system has been tested by 10 participants (25.8 ± 5.47 years; 74.69 ± 14.81 kg; 173.5 ± 9.52 cm) in DSP and LR training for four weeks. At the end of this period and according to the results of paired t-test analysis (p < 0.05) starting from the first week, it was observed that the participants trained more accurately and that they enhanced their motions by 58.08 ± 11.32% in LR training and 54.84 ± 12.72% in DSP training.

Guidelines for scoping manpower requirements of rule-based and frame-based expert systems are provided. An evolutionary methodology is presented that allows the estimation structure to increase as problem complexity increases. It is in the form of a six-month formula that enables organizations to get started in expert systems with minimal risk. The first project month is spent picking an expert system tool, isolating a problem, and writing a rudimentary project plan. The next two months are spent getting the first prototype up and running. The fourth month is divided between reviewing where you've been and where you're going. The fifth month is spent finishing the expert system, and the final month in completing system documentation and deployment.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>