Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Who's important? -- SUnSET: Synergistic Understanding of Stakeholder, Events and Time for Timeline Generation

Tiviatis Sim, Kaiwen Yang, Shen Xin, Kenji Kawaguchi

Abstract

As news reporting becomes increasingly global and decentralized online, tracking related events across multiple sources presents significant challenges. Existing news summarization methods typically utilizes Large Language Models and Graphical methods on article-based summaries. However, this is not effective since it only considers the textual content of similarly dated articles to understand the gist of the event. To counteract the lack of analysis on the parties involved, it is essential to come up with a novel framework to gauge the importance of stakeholders and the connection of related events through the relevant entities involved. Therefore, we present SUnSET: Synergistic Understanding of Stakeholder, Events and Time for the task of Timeline Summarization (TLS). We leverage powerful Large Language Models (LLMs) to build SET triplets and introduced the use of stakeholder-based ranking to construct a $Relevancy$ metric, which can be extended into general situations. Our experimental results outperform all prior baselines and emerged as the new State-of-the-Art, highlighting the impact of stakeholder information within news article.

Who's important? -- SUnSET: Synergistic Understanding of Stakeholder, Events and Time for Timeline Generation

Abstract

As news reporting becomes increasingly global and decentralized online, tracking related events across multiple sources presents significant challenges. Existing news summarization methods typically utilizes Large Language Models and Graphical methods on article-based summaries. However, this is not effective since it only considers the textual content of similarly dated articles to understand the gist of the event. To counteract the lack of analysis on the parties involved, it is essential to come up with a novel framework to gauge the importance of stakeholders and the connection of related events through the relevant entities involved. Therefore, we present SUnSET: Synergistic Understanding of Stakeholder, Events and Time for the task of Timeline Summarization (TLS). We leverage powerful Large Language Models (LLMs) to build SET triplets and introduced the use of stakeholder-based ranking to construct a metric, which can be extended into general situations. Our experimental results outperform all prior baselines and emerged as the new State-of-the-Art, highlighting the impact of stakeholder information within news article.

Paper Structure

This paper contains 26 sections, 12 equations, 9 figures, 12 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Works
  3. Methodology
  4. SET Generation
  5. Event Clustering
  6. Timeline Generation
  7. Experimental Setups
  8. Datasets.
  9. Benchmark.
  10. Models and Deployment.
  11. Metrics.
  12. Setups.
  13. Experimental Results
  14. Relevancy in Event Clustering
  15. Relevancy in Timeline Generation
  16. ...and 11 more sections

Figures (9)

  • Figure 1: An illustration of how SunSET generates a timeline for TLS through utilizing stakeholder information for relevance scoring.
  • Figure 2: Full SUnSET Framework for TLS. News articles generate SETs and Stakeholders undergo Coreference Resolution. Subsequently, events are clustered through Cosine Similarity and Relevance. The clusters will then be ranked while TextRank extracts the narrative for final timeline creation.
  • Figure 3: Effect of $\beta$ hyperparameter on $Rel$ in event clustering (TR) and $Rel$ in both event clustering and timeline generation (TR+$Rel$). The leftmost graph compares ROUGE-F1 values, the middle graph compares ROUGE-F2 values, and the rightmost graphs compare Date-F1 values.
  • Figure 4: Case-by-case representation of Penalty Behaviour
  • Figure 5: Representation of Penalty$_{IDF}$ Behaviour
  • ...and 4 more figures

Theorems & Definitions (1)

  • proof