Research and development of Community research in the field of AI and law should be pursued vigorously for several reasons. Community research can complement rules-based expert systems and improve their ability to think about legal predicates, solve problems effectively and explain their results. Community research can also contribute to the development of intelligent legal data retrieval systems and improve legal document creation programmes. Finally, in cognitive studies of various fields, it can model methods to transform poorly structured problems into better structured problems through case comparisons. On the negative side, critics argue that CBR`s main premise is based on anecdotal evidence and that matching items from one case to another can be complex and potentially lead to inaccuracies. However, recent work has improved the CBR through the use of a statistical framework. This makes it possible to make case-based predictions with a higher level of confidence. In everyday life, an auto mechanic who repairs an engine by recalling another car that had similar symptoms uses case-based reasoning. A lawyer advocating a particular outcome in a precedent-based trial, or a judge creating jurisprudence, uses case-based reasoning.
Even an engineer who copies functional elements of nature (practices biomimicry) treats nature as a problem-solving database. Case-based thinking is an important type of analogy solution. The first CBR legal model to deal with all these problems was incorporated into HYPO (Rissland and Ashley 1987, Ashley 1990), a computer program for analyzing cases and constructing legal arguments in the field of trade secrets. In the HYPO, legally relevant features, called dimensions, are used to index and retrieve cases. HYPO`s legal expertise includes, for each characteristic, which party to a dispute would benefit if the feature had certain values (e.g. a real value for predicates or a higher value for numerical features) and how the value of the characteristic relates to the underlying facts of a case. In the field of trade secrets, for example, such a characteristic is common knowledge, indicating that the alleged secret information is common knowledge in the industry. If a new case is submitted to the HYPO, the characteristics applicable to the new case are calculated. Next, HYPO constructs a “claims grid” – a similarity network that has the new case as the root node, and the previous cases that share a maximum set of characteristic values with the new case as immediate successors. These immediate successors are considered the most accurate cases.
Hammond, K. J. 1986. HEAD: A case-based planning model. In Proceedings AAAI-86, 267–271.Philadelphia: American Association for Artificial Intelligence. In CBR`s recent research on artificial intelligence (AI), five paradigmatic approaches have emerged: statistically oriented, model-based, planning/design-oriented, example-based, and adversarial or precedent-based. Paradigms differ in the assumptions they make about domain models, the extent to which they support the comparison of symbolic cases, and the types of reasoning for which they use cases. Case-based reasoning (CBR) is an experience-based approach to solving new problems by adapting previously successful solutions to similar problems. CBR deals with memory, learning, planning and problem solving, providing a foundation for a new technology of intelligent computer systems that can solve problems and adapt to new situations. In RCC, the “smart” reuse of knowledge from problems or cases that have already been solved is based on the premise that the more similar two problems are, the more similar their solutions will be. Subsequent research expanded the use of HYPO-style CBR in several ways.
In the GREBE system (Branting 1991), a ratio decidendi presentation of cases was used to identify those parts of the stocks of a previous case that were relevant to a new case, even if other parts of the previous case were not applicable. Cranning`s model addressed a well-known weakness of the purely characteristic-based approach to RCC: the inability to present the reasoning that relates characteristics to the outcomes of a case.