case-based reasoning

A problem-solving and learning paradigm in which a system handles a new problem by retrieving similar past cases from a case base, reusing their solutions, revising them to fit the new situation and retaining the result as a new case — the 'four Rs' cycle. Retrieval depends on a similarity measure over case features, and adaptation may use rules, analogy or human input. Because knowledge is stored as concrete episodes rather than general rules, case-based systems ease the knowledge acquisition bottleneckknowledge acquisition bottleneck, n.: The problem that getting knowledge out of human experts and into a computer by hand is so slow and costly that it limited how far classic expert systems could go., learn incrementally and can justify answers by pointing to precedents; their weaknesses are dependence on a good similarity measure and the difficulty of automatic adaptation. Applications include help-desk diagnosis, legal argument, engineering design and recipe generation. Retrieval-augmented generationretrieval-augmented generation, n.: Giving an AI an open-book exam: before answering, it looks up relevant passages in a set of documents and bases its reply on them, often citing where each point came from. and the selection of examples for few-shotfew-shot learning, n.: Getting an AI system to handle a new task or recognize a new kind of thing from just a handful of examples, the way a person can learn a new word after hearing it once or twice. prompting are often compared to it.