Summary of Efficient Long-distance Latent Relation-aware Graph Neural Network For Multi-modal Emotion Recognition in Conversations, by Yuntao Shou et al.
Efficient Long-distance Latent Relation-aware Graph Neural Network for Multi-modal Emotion Recognition in Conversations
by Yuntao Shou, Wei Ai, Jiayi Du, Tao Meng, Haiyan Liu, Nan Yin
First submitted to arxiv on: 27 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
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 proposes an Efficient Long-distance Latent Relation-aware Graph Neural Network (ELR-GNN) for multi-modal emotion recognition in conversations. Existing methods rely on graph neural networks to model conversational relationships, but struggle with capturing dependencies between long-distance utterances. To address this, the authors first extract text, video, and audio features using Bi-LSTM and then construct a conversational emotion interaction graph. They use a dilated generalized forward push algorithm to precompute emotional propagation and design an emotional relation-aware operator to capture semantic associations between utterances. Early fusion and adaptive late fusion mechanisms are combined to fuse latent dependency information with speaker relationship and context. The proposed model achieves state-of-the-art performance on IEMOCAP and MELD benchmark datasets, while reducing running times by 52% and 35%, respectively. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to understand emotions in conversations. It’s called multi-modal emotion recognition in conversation (MERC). Right now, computers are not very good at understanding how people feel when they talk. The authors of this paper want to change that by making a special kind of computer program that can analyze conversations and figure out what people are feeling. They’re using something called graph neural networks, which is like a map that shows how different parts of the conversation relate to each other. This helps the program understand things that are happening far apart in the conversation. The new program is faster and better than old programs at doing this task. |
Keywords
* Artificial intelligence * Gnn * Graph neural network * Lstm * Multi modal
