研究人员开发了一个名为ML-SAN的新模型，通过考虑个体表达差异来改进对话中的情感识别。这个多级说话人自适应网络使用一个三阶段过程来校准输入特征，根据说话人身份调整模态信任度，并在潜在空间中保持说话人一致性。在MELD和IEMOCAP数据集上的测试表明，ML-SAN表现更好，尤其是在不太常见的感情类别和多样化的说话人方面。 AI
影响 通过适应个体说话人表达风格来改进多模态情感识别，增强机器共情能力。
排序理由 这是一篇介绍情感识别新模型的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
To establish empathy with machines, it is essential to fully understand human emotional changes. However, research in multimodal emotion recognition often overlooks one problem: individual expressive traits vary significantly, which means that different people may express emotion…
To establish empathy with machines, it is essential to fully understand human emotional changes. However, research in multimodal emotion recognition often overlooks one problem: individual expressive traits vary significantly, which means that different people may express emotion…