一种用于CBR推理机的案例学习算法研究*

案例学习是CBR(Case-Based Reasoning)推理机的重要环节,但由于案例的多样性以及对领域的依赖性,导致CBR系统中案例自动生成困难的问题。针对这一问题,本文提出将seq2seq(Sequence-to-Sequence)模型用于案例学习,通过seq2seq模型自动生成案例,引入attention机制,提高seq2seq模型生成案例的效果,并利用潜在语义分析LSA(Latent Semantic Analysis)对网络爬取语料库进行筛选,利用过滤后的语料库对模型进行训练,提出一种基于三元组的评估方法,对生成案例进行评估和存储,从而实现CBR推理机的自主学习。最后将改进的案例学习系统应用到实际的智能机器人上进行验证,测试结果表明该方法具有可行性,且能够有效提高机器人的智能性及易用性。

Case study algorithm of CBR reasoning machine

Case study is an important part of CBR reasoning machine, but because of the diversity of cases and the dependence on the field, the cases automatically generate is difficult in CBR system. To solve this problem, this paper proposes to use the seq2seq (Sequence-to-Sequence) model for case study. Seq2seq model is used to automatically generate cases and an attention mechanism is used to make the model generate a better case. The training corpus of the model is crawled from the network and filtered by the LSA algorithm. In order to achieve the self-learning of CBR reasoning machine, this paper also proposes a case evaluation algorithm based on triad thought. Finally, the improved case learning system is applied to the actual intelligent robot. The test results show that the method is feasible and can improve the intelligence and ease of use of the robot effectively.