AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Span-Based Event Coreference Resolution

Jing Lu, Vincent Ng

AAAI 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 18 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use span-based joint trigger detection and event coreference instead of the compared prior resolverKBP 2017 event coreference · KBP 2017 testAVG-F: 37.9 -> 43.8Improved
Add a feature indicating whether candidate event mentions share their gold subtype during training (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 44.6Improved
Add a feature indicating whether candidate event mentions share their predicted subtype (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 44.1Improved
Hard-filter candidate antecedents with a different event subtype (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 43.9No clear effect
Add soft subtype-consistency constraints to candidate link scores (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 44.5Improved
Use trigger-detection/coreference consistency methods (ablation)Span-based event coreference · KBP 2017 testAVG-FImproved
Add independently computed entity coreference information as features (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 44.9Improved
Add independently computed entity coreference information as a hard constraint (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 45.1Improved
Jointly train entity coreference with event coreference and use entity-coreference features (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 46.4Improved
Jointly train entity coreference and use its output as a hard constraint (ablation)Span-based event coreference · KBP 2017 testAVG-F: 43.8 -> 46.7Improved
Add entity coreference information using any evaluated feature or constraint combination (ablation)Span-based event coreference · KBP 2017 testAVG-FImproved
Use jointly trained rather than independently trained entity coreference information (ablation)Span-based event coreference with entity information · KBP 2017 testAVG-FImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy