The Data Mining Lab (DM Lab) of SRCB participated in evaluation workshop of the Text Analysis Conference (TAC) 2017 and got good performance, ranking Top 1 for 3 asks and ranking Top 3 for left 2 tasks of Event Nugget Detection and Coreference.
The Text Analysis Conference is organized by the U.S. National Institute of Standards and Technology (NIST), sponsored by the U.S. Department of Defense. It is a series of evaluation workshops organized to encourage research in Natural Language Processing and related applications, by providing a large test collection, common evaluation procedures, and a forum for organizations to share their results.
TAC 2017 has released six tracks in domains of Knowledge Base Population (KBP) and Adverse Drug Reaction Extraction (ADR), attracting top researchers from BBN, IBM Research, RPI, Oracle, CMU, Tencent, Alibaba, NEC, LDC, Columbia, USC ISI, UTHealth, UND and other well-known international academic and industrial research institutions nearly sixty teams to participate.
DM Lab participated in the evaluation of KBP, and ranked No.1 on 3 tasks of Event Nugget Detection, Event Type and Subtypes, and Event Coreference. Especially, on 2 tasks of Event Nugget Detection and Event Type and Subtypes, the performance is better than the second position by 7.32% and 5.82% respectively.
The technology related to knowledge graph populationis the key research direction of DM Lab in recent years. As the fundamental of future AI, it will provide strong support for upgrading RICOH products.