Summary of Are Llms Good Annotators For Discourse-level Event Relation Extraction?, by Kangda Wei et al.
Are LLMs Good Annotators for Discourse-level Event Relation Extraction?
by Kangda Wei, Aayush Gautam, Ruihong Huang
First submitted to arxiv on: 28 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper investigates the effectiveness of Large Language Models (LLMs) in extracting event relations from lengthy documents, focusing on discourse-level event relation extraction (ERE) tasks. The study finds that LLMs, despite their proficiency in natural language processing tasks, underperform compared to a supervised learning baseline. Although Supervised Fine-Tuning (SFT) can improve LLM performance, it does not scale well compared to the smaller supervised baseline model. The paper highlights several weaknesses of LLMs in ERE tasks, including fabricating event mentions, failing to capture transitivity rules, detect long-distance relations, or comprehend contexts with dense event mentions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how well Large Language Models (LLMs) can do when we ask them to find relationships between events in long documents. LLMs are good at many language tasks, but they don’t do as well on this specific task called discourse-level event relation extraction (ERE). The study shows that even with some help from training, the LLMs still don’t do as well as a simpler model that was trained just for this task. This paper also points out some weaknesses in how LLMs work when it comes to finding these relationships. |
Keywords
» Artificial intelligence » Discourse » Fine tuning » Natural language processing » Supervised
