New York University 2014 Knowledge Base Population Systems

New York University (NYU) participated in three tracks of the 2014 TAC-KBP eval- uation: English Slot Filling, Cold Start and Entity Discovery and Linking. While this year is the first time and second time we participated in entity discovery and link- ing (EDL) and cold start respectively, we have been working on the slot filling task for several years. With additional devel- opment time this year, our cold start sys- tem has borrowed more technologies from slot filling and gained significant improve- ment. In both slot filling and cold start systems, we find that besides incorporat- ing new modules to expand the coverage, the systems greatly benefit from our ef- fort to renovate some of the basic mod- ules. Regarding entity discovery and link- ing, we develop a pipelined system with several components in which we introduce a variation of the PageRank algorithm to improve collaborative candidate ranking.