New adapter DiaRelay enhances LLMs for emotion recognition in conversations

Researchers have developed DiaRelay, a novel adapter for Large Language Models (LLMs) designed to improve Emotion Recognition in Conversation (ERC). Unlike existing methods that use fixed context windows or re-encode entire histories, DiaRelay employs a constant-size memory to selectively aggregate and propagate historical dialogue evidence. This approach allows for better tracking of subtle emotional cues across distant turns without increasing computational costs or context length. Experiments on MELD and IEMOCAP datasets demonstrate that DiaRelay achieves state-of-the-art results with a minimal number of trainable parameters. AI

影响 This research could lead to more nuanced and context-aware AI systems for understanding and generating human-like dialogue.

排序理由 The cluster describes a new research paper detailing a novel method for emotion recognition in conversations using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

• DiaRelay
• Emotion Recognition in Conversation (ERC)
• IEMOCAP
• Large Language Models (LLMs)
• LoRA
• MELD

• 论文
• 模型发布

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →