Abstract

We participated in the SM-KBP 2020 evaluation using our dockerlized GAIA system, an end-toend knowledge extraction, grounding, inference, clustering, temporal tracking and hypothesis generation system, as shown in Figure 1. Our TA1 system achieves top performance at both intrinsic evaluation and extrinsic evaluation through TA2 and TA3. In the past year, we integrate the following innovations:
• Multilingual Joint Information Extraction with Global Knowledge: We propose an end-to-end neural model ONEIE to extract entities, relations and events jointly in a language independent fashion. Existing joint neural models for Information Extraction (IE) use local task-specific classifiers to predict labels for individual instances (eg, trigger, relation) regardless of their interactions. For example, a VICTIM of a DIE event is likely to be a VICTIM of an ATTACK event in the same sentence. Our model can capture such cross-subtask and cross-instance inter-dependencies, we extract the globally optimal information network by considering the inter-dependency among nodes and edges. At the decoding stage, we incorporate global features to capture the cross-subtask

Date

2020

Authors

Manling Li, Ying Lin, Tuan Manh Lai, Xiaoman Pan, Haoyang Wen, Sha Li, Zhenhailong Wang, Pengfei Yu, Lifu Huang, Di Lu, Qingyun Wang, Haoran Zhang, Qi Zeng, Chi Han, Zixuan Zhang, Yujia Qin, Xiaodan Hu, Nikolaus Parulian, Daniel Campos, Heng Ji, Brian Chen, Xudong Lin, Alireza Zareian, Amith Ananthram, Emily Allaway, Shih-Fu Chang, Kathleen McKeown, Yixiang Yao, Michael Spector, Mitchell DeHaven, Daniel Napierski, Marjorie Freedman, Pedro Szekely, Haidong Zhu, Ram Nevatia, Yang Bai, Yifan Wang, Ali Sadeghian, Haodi Ma, Daisy Zhe Wang

Journal

Proceedings of Thirteenth Text Analysis Conference (TAC 2020)