Summary of Causal Discovery Inspired Unsupervised Domain Adaptation For Emotion-cause Pair Extraction, by Yuncheng Hua et al.
Causal Discovery Inspired Unsupervised Domain Adaptation for Emotion-Cause Pair Extraction
by Yuncheng Hua, Yujin Huang, Shuo Huang, Tao Feng, Lizhen Qu, Chris Bain, Richard Bassed, Gholamreza Haffari
First submitted to arxiv on: 18 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 This paper tackles emotion-cause pair extraction in unsupervised domain adaptation, where distributions of events causing emotions differ dramatically between source and target domains. Despite overlapped emotional expressions, the task is challenging. The authors propose a novel deep latent model based on variational autoencoders (VAEs) that captures underlying structures and utilizes easily transferable emotion knowledge to link event distributions across domains. To facilitate knowledge transfer, they also propose a variational posterior regularization technique to disentangle latent representations of emotions from events and mitigate spurious correlations. The authors demonstrate their model outperforms the strongest baseline by approximately 11.05% on a Chinese benchmark and 2.45% on an English benchmark in terms of weighted-average F1 score, using CAREL-VAE. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps computers understand why people feel certain emotions. It’s like trying to figure out what makes someone happy or sad just by looking at their words. The problem is tricky because the things that happen that make us feel a certain way are very different in different places, even though we might say the same things to express our feelings. To solve this problem, the authors create a new kind of computer model that uses patterns it finds in data to help computers understand emotions and connect them to what caused those emotions. They test their model on two big datasets and show it works better than other methods. |
Keywords
» Artificial intelligence » Domain adaptation » F1 score » Regularization » Unsupervised
