Paper: Evolutionary Hierarchical Dirichlet Process for Timeline Summarization

ACL ID P13-2099
Title Evolutionary Hierarchical Dirichlet Process for Timeline Summarization
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2013
Authors

Timeline summarization aims at generat- ing concise summaries and giving read- ers a faster and better access to under- stand the evolution of news. It is a new challenge which combines salience rank- ing problem with novelty detection. Pre- vious researches in this field seldom ex- plore the evolutionary pattern of topics such as birth, splitting, merging, develop- ing and death. In this paper, we develop a novel model called Evolutionary Hier- archical Dirichlet Process(EHDP) to cap- ture the topic evolution pattern in time- line summarization. In EHDP, time vary- ing information is formulated as a series of HDPs by considering time-dependent information. Experiments on 6 different datasets which contain 3156 documents demonstrates the good performance of our system with regard to ROUGE ...