Definition
Expert Systems are a type of artificial intelligence developed primarily in the 1980s that use rule-based logic to solve complex problems in specific domains. They rely on a “knowledge base” of facts and rules elicited from human experts and an “inference engine” to apply those rules to specific cases.
Why It Matters
Understanding Expert Systems is vital because they represent the first major attempt to institutionalize human knowledge in software. Their ultimate “brittleness” and the subsequent AI Winter serve as a cautionary tale for modern AI development, highlighting the danger of top-down, rigid logic compared to the adaptive, bottom-up learning of neural networks. They remain relevant as high-reliability components in narrow, high-stakes domains where “black box” decisions are unacceptable.
Core Concepts
- Rule-Based Reasoning: Logic follows “IF [condition] THEN [action]” patterns.
- Hand-Coded Knowledge: Knowledge is painstakingly elicited from domain experts and translated into formal code by “knowledge engineers.”
- Domain Specificity: They excel in narrow fields (e.g., medical diagnosis, mineral exploration) but lack general “common sense” or ability to handle tasks outside their pre-defined rules.
- Brittleness: A hallmark of GOFAI (Good Old-Fashioned AI). If a user provides input slightly outside the system’s narrow assumptions, the system often crashes or produces nonsensical results.
- Maintenance Bottleneck: Large systems proved expensive to update, validate, and keep current as human knowledge evolved.