Summary of Ttm-re: Memory-augmented Document-level Relation Extraction, by Chufan Gao et al.
TTM-RE: Memory-Augmented Document-Level Relation Extraction
by Chufan Gao, Xuan Wang, Jimeng Sun
First submitted to arxiv on: 9 Jun 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 Medium Difficulty summary: This paper proposes a novel approach called TTM-RE for document-level relation extraction, which leverages large amounts of noisy training data. Unlike previous methods that struggle with varying noise levels, TTM-RE integrates a trainable memory module, the Token Turing Machine, with a noisy-robust loss function to improve performance. The approach achieves state-of-the-art results on the ReDocRED benchmark dataset, outperforming previous methods by over 3% in terms of F1 score. Additionally, ablation studies demonstrate the effectiveness of TTM-RE in other domains and under highly unlabeled settings. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: This research is about finding relationships between things mentioned in documents. Right now, most methods don’t work well when they’re trained on large amounts of data that are a bit noisy or imperfect. The new approach, called TTM-RE, is designed to overcome this limitation by using a special kind of memory and a way to adjust for the noise in the training data. It does much better than previous methods on a test dataset, which shows it’s effective at finding relationships between things. |
Keywords
» Artificial intelligence » F1 score » Loss function » Token
