What is case-based reasoning (CBR)? 基于案例的推理

这个方向其实挺小众的，然后其实也很直觉，所以你可能其实一直在用这个技术，只不过你不知道这个技术有一个专业的名字《基于案例的推理》而已。

有人说，不知道名字有啥的，反正大家都会用，不需要学它叫什么名字。话虽如此，但是如果你说你的这种技术大有来头，等到你汇报给别人听的时候，或者写论文的时候，别人就会觉得你的东西逼格高，而不是干巴巴的说，我们提出了一个方法，这个方法非常直觉。

What is case-based reasoning (CBR)?

Case-based reasoning (CBR) is an experience-based approach to solving new problems by adapting previously successful solutions to similar problems. Addressing memory, learning, planning and problem solving, CBR provides a foundation for a new type of intelligent computer system that can solve problems and adapt to new situations. In CBR, the “intelligent” reuse of knowledge from already-solved problems or cases relies on the premise that the more similar two problems are, the more similar their solutions will be.

CBR developed its roots in the work of artificial intelligence (AI) theorist and cognitive psychologist, Roger Schank, and his students at Yale in the late 20th century. These researchers studied the problem-solving ability of humans and found that most people assemble solutions based on earlier experiences with similar situations.

With applications spanning fields ranging from machine learning to medicine to law, CBR is accomplished by gathering case histories and implemented by identifying significant features that describe a case. CBR systems can “learn” by acquiring new knowledge as cases. This, along with the application of database techniques, makes it easier to maintain large volumes of information.

4-step process for CBR

In general, the case-based reasoning process entails four steps:

1.Retrieval. Gathering from memory an experience closest to the current problem.
2.Reuse. Suggesting a solution based on the experience and adapting it to meet the demands of the new situation.
3.Revision. Evaluating the use of the solution in the new context.
4.Retaining. Storing this new problem-solving method in the memory system.

Comparison to other techniques

Case-based reasoning has several differences from other AI approaches, such as knowledge-based systems. Rather than relying completely on general knowledge of a problem domain or making associations along generalized relationships between problem descriptors and conclusions, CBR employs the specific knowledge of previously experienced, concrete problem situations. CBR also offers incremental, sustained learning in that each time a problem is solved a new experience is retained and can be applied for future problems.

Advantages and disadvantages of CBR

Scientists cite both advantages and disadvantages to CBR. On the plus side, remembering past experiences helps learners avoid repeating previous mistakes, and the reasoner can discern what features of a problem are significant enough to focus on.

Further, CBR is intuitive because it reflects how people work. Because no knowledge must be elicited to create rules or methods, development is easier. Another benefit is that systems learn by acquiring new cases through use, and this makes maintenance easier.

CBR also enables the reasoner to propose solutions to problems quickly. The reasoner can propose solutions in areas that he or she does not fully understand, evaluate solutions when no algorithmic method is available and interpret open-ended and ill-defined concepts.

On the negative side, critics claim that the main premise of CBR is based on anecdotal evidence and that adapting the elements of one case to another may be complex and potentially lead to inaccuracies. However, CBR has more recently been enhanced using statistical frameworks. This makes it possible to produce case-based predictions with a higher level of confidence.