How Much Do You Know About Expert Systems?

Expert Systems: A Summary of History, Architecture, Advantages, and Disadvantages

• Expert systems are computer systems that emulate the decision-making ability of a human expert.

• Expert systems are designed to solve complex problems by reasoning through bodies of knowledge represented mainly as if-then rules.

• The first expert systems were created in the 1970s and proliferated in the 1980s, becoming the first successful forms of artificial intelligence software.

• Expert systems are divided into two subsystems: the inference engine and the knowledge base.

• The inference engine applies rules to known facts to deduce new facts and can include explanation and debugging abilities.

• The history of expert systems dates back to the late 1950s when researchers started experimenting with computer technology to emulate human decision-making, particularly in the medical field.

• Expert systems were formally introduced around 1965 by the Stanford Heuristic Programming Project.

• In the 1980s, expert systems proliferated with two-thirds of Fortune 500 companies applying the technology in daily business activities.

• In the 1990s and beyond, the term expert system mostly dropped from the IT lexicon, and Rule-Based Systems became a more popular term.

• Modern expert systems incorporate new knowledge more easily, can generalize from existing knowledge better, and can deal with vast amounts of complex data.

• Expert systems have an explicit knowledge representation, which makes maintenance easier and enables rapid prototyping.

• The most common disadvantage cited for expert systems in the academic literature is the knowledge acquisition problem.Challenges and Disadvantages of Expert Systems

• The performance of early expert systems was a major problem due to the use of interpreted code expressions instead of compiled languages, resulting in slower processing.

• Integration with legacy systems and databases was difficult for early expert systems due to the use of programming languages and hardware platforms that were not compatible with most corporate IT environments.

• The processing complexity increases as the size of the knowledge base increases, making it difficult for the inference engine to process a huge number of rules to reach a decision.

• Consistency verification of decision rules with each other becomes a challenge with too many rules, leading to a satisfiability (SAT) formulation, which is a well-known NP-complete problem.

• Prioritizing the use of rules to operate more efficiently and resolving ambiguities are also challenging.

• Overfitting and overgeneralization effects are also problems when trying to generalize known facts to other cases not explicitly described in the knowledge base.

• Updating the knowledge base quickly and effectively and adding a new piece of knowledge among many rules is challenging.

• New approaches based on machine learning techniques and feedback mechanisms are required instead of rule-based technologies.

• Expert systems in medicine face challenges related to big data, existing regulations, healthcare practice, various algorithmic issues, and system assessment.

• Expert systems applications are divided into 10 categories, including medical diagnosis systems, salespeople configuring computers, and dam safety monitoring.

• Hearsay, an early attempt at solving voice recognition through an expert systems approach, was not successful.

• GARVAN-ES1 was one of the first medical expert systems to go into routine clinical use internationally, and the first expert system to be used for diagnosis daily in Australia.

• Mistral is an expert system to monitor dam safety, developed in the 1990s by Ismes (Italy), and is still operational 24/7/365 on the Ridracoli Dam (Italy).