聚焦情感热点：面向对话情感识别的多模态局部-全局融合与跨模态对齐
Centering Emotion Hotspots: Multimodal Local-Global Fusion and Cross-Modal Alignment for Emotion Recognition in Conversations
AI总结：
针对对话情感识别中证据稀疏与模态异步问题，提出以情感热点为中心的统一模型，通过热点门控融合与路由混合对齐器实现局部-全局融合及跨模态对齐，在标准基准上超越强基线。
摘要：
对话情感识别（ERC）的难点在于判别性证据稀疏、局部化，且在不同模态间往往异步。我们将ERC聚焦于情感热点，并提出一个统一模型，该模型在文本、音频和视频中检测每个话语的热点，通过热点门控融合（Hotspot-Gated Fusion）将其与全局特征融合，并使用路由混合对齐器（routed Mixture-of-Aligners）对齐模态；跨模态图编码对话结构。这种设计将建模聚焦于显著片段，缓解错位问题，并保留上下文。在标准ERC基准上的实验表明，相较于强基线模型，该模型持续取得提升，消融实验证实了HGF和MoA的贡献。我们的结果指向一种以热点为中心的观点，可为未来的多模态学习提供参考，为ERC中的模态融合提供新视角。
英文摘要：
Emotion Recognition in Conversations (ERC) is hard because discriminative evidence is sparse, localized, and often asynchronous across modalities. We center ERC on emotion hotspots and present a unified model that detects per-utterance hotspots in text, audio, and video, fuses them with global features via Hotspot-Gated Fusion, and aligns modalities using a routed Mixture-of-Aligners; a cross-modal graph encodes conversational structure. This design focuses modeling on salient spans, mitigates misalignment, and preserves context. Experiments on standard ERC benchmarks show consistent gains over strong baselines, with ablations confirming the contributions of HGF and MoA. Our results point to a hotspot-centric view that can inform future multimodal learning, offering a new perspective on modality fusion in ERC.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
