计算机软件新技术国家重点实验室 

摘 要：

      Semantic spaces are vector-space representations of entities that are mainly used as geometric models of semantic relatedness. Semantic spaces have been widely studied in the cognitive science literature to model phenomena such as categorization and induction. They have traditionally been estimated by applying multi-dimensional scaling to human similarity judgments, although in computer science they are usually estimated from bag-of-words representations, using variants of word embedding models. Such learned semantic spaces are commonly used to support inductive inferences, e.g. to predict missing type information in knowledge bases. In this talk, I will give an overview of some approaches for learning semantic spaces and how to use them for reasoning.

     Zied Bouraoui is an associate professor (Maître de conférences, in French) at Artois University, France. His research interests are at the intersection of Machine Learning (ML), Natural Language Processing (NLP) and Knowledge Representation (KR). He is currently working on automated knowledge base completion and the development of flexible methods for reasoning to overcome some of the limitations of logical deduction. He is also interested in the development of methods for reasoning under uncertainty and/or inconsistency, belief merging/revision and query answering.

报告人:Zied Bouraoui

时间：3月21日 12:45-13:45