Genetic Programming for Feature Selection and Question-Answer Ranking in IBM Watson

摘要: IBM Watson is an intelligent open-domain question answering system capable of finding correct answers to natural language questions in real-time. uses machine learning over a large heterogeneous feature set derived from many distinct processing algorithms identify answers. This paper develops Genetic Programming (GP) approach for selection by evolving ranking functions order candidate generated Watson. We leverage GP’s automatic mechanisms Watson’s key features through the process. Our experiments show that GP can evolve relatively simple use much fewer original achieve comparable performances methodology aid implementers better components otherwise and complex development, troubleshooting, and/or customer or domain-specific enhancements.