Files in this item

FilesDescriptionFormat

application/pdf

application/pdfZENG-THESIS-2021.pdf (1MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:Event network embedding
Author(s):Zeng, Qi
Advisor(s):Ji, Heng
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):natural language processing
Abstract:Current methods for event representation ignore related events in a corpus-level global context. For a deep and comprehensive understanding of complex events, we introduce a new task, Event Network Embedding, which aims to represent events by capturing the connections among events. We propose a novel framework, Global Event Network Embedding (GENE), that encodes the event network with a multi-view graph encoder while preserving the graph topology and node semantics. The graph encoder is trained by minimizing both structural and semantic losses. We develop a new series of structured probing tasks, and show that our approach effectively outperforms baseline models on node typing, argument role classification, and event coreference resolution. As a direct application, We introduce a new task, Unsupervised Event Schema Graph Matching, which aims to align event instance graphs and event schema graphs by finding node correspondence. We develop the first benchmark and collect a dataset of 3,740 event instance graphs from IED-scenario news articles, 75 of which are paired with 4 event schema graphs by human annotation. Our analysis on this task sheds light on the shortcomings of current state-of-the-art models on this event understanding task.
Issue Date:2021-12-09
Type:Thesis
URI:http://hdl.handle.net/2142/114027
Rights Information:Copyright 2021 Qi Zeng
Date Available in IDEALS:2022-04-29
Date Deposited:2021-12


This item appears in the following Collection(s)

Item Statistics