Papers by I-Ta Lee

4 papers
Modeling Human Mental States with an Entity-based Narrative Graph (2021.naacl-main)

Copied to clipboard

Challenge: Understanding narrative text requires capturing characters’ motivations, goals, and mental states.
Approach: They propose an Entity-based Narrative Graph (ENG) to model the internal-states of characters in a story and evaluate it on two narrative understanding tasks.
Outcome: The proposed model is based on two narrative understanding tasks: predicting character mental states, and desire fulfillment.
A Structured Clustering Approach for Inducing Media Narratives (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to modeling media narratives miss subtle narrative patterns through coarse-grained analysis or require domain-specific taxonomies that limit scalability.
Approach: They propose a framework for inducing rich narrative schemas by jointly modeling events and characters via structured clustering.
Outcome: The proposed framework produces explainable narrative schemas that align with established framing theory while scaling to large corpora without exhaustive manual annotation.
Multi-Relational Script Learning for Discourse Relations (P19-1)

Copied to clipboard

Challenge: Existing script knowledge models only represent a single event relationship, co-occurrence . this is coarse for commonsense, which should account for fine-grained relationships .
Approach: They propose to view learning event embedding as a multi-relational problem . they model a rich set of event relations derived from the Penn Discourse Tree Bank .
Outcome: The proposed model captures different aspects of event pairs, including cause and contrast.
Weakly-Supervised Modeling of Contextualized Event Embedding for Discourse Relations (2020.findings-emnlp)

Copied to clipboard

Challenge: Structured knowledge representations capture temporal relations between events to describe human-level representations of common scenarios.
Approach: They propose to represent narrative graphs and learn contextualized event representations over them using a relational graph neural network model.
Outcome: The proposed model improves performance when learning script knowledge without supervision and provides a better representation for the implicit discourse sense classification task.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations