Universal Schema for Slot Filling and Cold Start: UMass IESL at TACKBP 2013
نویسندگان
چکیده
We employ universal schema for slot filling and cold start. In universal schema, we allow each surface pattern from raw text, and each type defined in ontology, i.e. TACKBP slots to represent relations. And we use matrix factorization to discover implications among surface patterns and target slots. First, we identify mentions of entities from the whole text corpus and extract relations between entity pairs to construct a knowledge base. Finally, we query this knowledge base to produce our submissions for slot filling and cold start.
منابع مشابه
Building Knowledge Bases with Universal Schema: Cold Start and Slot-Filling Approaches
We compare the performance of two different relation prediction architectures based on the same relation predictors. The knowledge base construction architecture builds a complete knowledge base for the entire corpus, and commits to entity linking and clustering decisions ahead of time. The query-driven slot filling architecture can make entity expansion and retrieval decisions on the fly, and ...
متن کاملNew York University 2014 Knowledge Base Population Systems
New York University (NYU) participated in three tracks of the 2014 TAC-KBP evaluation: 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 linking (EDL) and cold start respectively, we have been working on the slot filling task for several years. With additional development time this year, o...
متن کاملOff to a Cold Start: New York University's 2013 Knowledge Base Population Systems
New York University submitted two runs this year: a KBP Slot Filling run and a Cold Start run. The Slot Filling run used the same system as last year (Min et al. 2012), with only the minimal changes needed to address the revised specifications. These included changes in the provenance information and collapsing employee_of and member_of slots. We included the confidence estimation step (Li and ...
متن کاملStanford’s 2014 Slot Filling Systems
We describe Stanford’s entry in the TACKBP 2014 Slot Filling challenge. We submitted two broad approaches to Slot Filling: one based on the DeepDive framework (Niu et al., 2012), and another based on the multi-instance multi-label relation extractor of Surdeanu et al. (2012). In addition, we evaluate the impact of learned and hard-coded patterns on performance for slot filling, and the impact o...
متن کاملA distant supervised learning system for the TAC-KBP Slot Filling and Temporal Slot Filling Tasks
This paper describes the system implemented by the NLP GROUP AT UNED for our first participation in the Knowledge Base Population at the Text Analysis Conference (TACKBP). For this Slot Filling Task, our approach was to design a distant supervised learning system, which was then specialized for the Regular Slot Filling and Full Temporal Slot Filling subtasks. From the initial Knowledge Base and...
متن کامل