CASE-BASED CBR - CAPTURING AND REUSING REASONING ABOUT CASE ADAPTATION

Case-based reasoning (CBR) solves new problems by retrieving solutions to similar prior problems and adapting them to fit new needs. Progres s in retrieval methods for CBR has resulted in a flourishing technolog y of case-based ''aiding systems'' that support human problem-solving by automatically providing the user with relevant cases. However, deve loping effective methods for automated case adaptation remains a centr al research challenge for the field. This article proposes alleviating the adaptation problem by using a case-based adaptation component to capture and reuse the reasoning underlying successful adaptations. Mor e generally, it illustrates the potential far case-based intelligent c omponents (Riesbeck, 1996) within CBR systems, presents specific princ iples and observations concerning their application to autonomous and interactive case adaptation, and suggests their potential role in impr oving similarity assessment and case retrieval. In addition, it propos es that adaptation learning helps to address the increasingly importan t problem of case-base maintenance, by enabling a ''lazy updating'' of the case base as new knowledge is acquired.