Related Experiment Videos
Hybrid expert system for decision supporting in the medical area: complexity and cognitive computing
L M Brasil1, F M de Azevedo, J M Barreto
1Informatics Laboratory in Health (LABIS), Nucleus of Studies and Technology in Biomedical Engineering (NETEB), Federal University of Paraíba, João Pessoa, Brazil. lmb@neteb.ufpb.br
International Journal of Medical Informatics
|August 24, 2001
Summary
This study introduces a hybrid expert system (HES) integrating cognitive computing to address AI challenges. It offers solutions for knowledge elicitation, reasoning representation, and explains AI conclusions, demonstrated with an epileptic crisis case study.
Area of Science:
- Artificial Intelligence
- Cognitive Computing
- Expert Systems
Background:
- Expert systems face challenges in knowledge elicitation, knowledge representation, and explaining AI reasoning.
- Connectionist approaches present difficulties in determining optimal network topology and neuron count.
- Obtaining clear explanations for AI-driven conclusions remains a significant hurdle.
Purpose of the Study:
- To propose a hybrid expert system (HES) that minimizes complexity issues in artificial intelligence.
- To integrate cognitive computing to enhance the capabilities of expert systems.
- To develop algorithms for training fuzzy neural networks and generating explanations for their conclusions.
Main Methods:
- Development of a hybrid expert system (HES) incorporating cognitive computing principles.
- Application of two novel algorithms: one for fuzzy neural network training, another for explanation generation.
- Utilizing a case study involving epileptic crisis for system validation and simulation.
Main Results:
- The proposed HES effectively addresses key AI complexities, including knowledge elicitation and reasoning transparency.
- The developed algorithms facilitate efficient training of fuzzy neural networks and provide comprehensible explanations.
- Simulations on epileptic crisis data demonstrate the system's practical applicability and performance.
Conclusions:
- The hybrid expert system offers a robust solution to persistent challenges in artificial intelligence.
- Integration of cognitive computing and specialized algorithms enhances the explainability and efficiency of AI systems.
- The system shows promise for real-world applications, as evidenced by the epileptic crisis case study.