SemEval-2024 Task 3: Multimodal Emotion Cause Analysis in Conversations

理解情感是类人工智能中不可或缺的组成部分，因为情感极大地影响人类的认知、决策和社交互动。除了在对话中识别情感，识别对话中个体情感状态背后的潜在原因，在许多应用场景中也非常重要。我们组织了SemEval-2024任务3，名为“对话中的多模态情感原因分析”，旨在从对话中提取所有情感及其对应原因的配对。在不同的模态设置下，它包括两个子任务：对话文本情感-原因配对提取（TECPE）和对话多模态情感-原因配对提取（MECPE）。这项共享任务吸引了143个注册和216个成功提交。在本文中，我们介绍了任务、数据集和评估设置，总结了顶尖团队的系统，并讨论了参与者的发现。

SemEval-2024 Task 3, named Multimodal Emotion Cause Analysis in Conversations, aims to extract all pairs of emotions and their corresponding causes from conversations. The task addresses the need for emotion recognition and cause analysis in human-like AI, which greatly influences decision making and social interactions.

The task consists of two subtasks: Textual Emotion-Cause Pair Extraction in Conversations (TECPE) and Multimodal Emotion-Cause Pair Extraction in Conversations (MECPE). The shared task attracted 143 registrations and 216 successful submissions.

The paper introduces the task, dataset and evaluation settings, summarizes the systems of the top teams, and discusses the findings of the participants. The dataset used for the task is provided and the top-performing systems are open-sourced. The task provides a platform for further research in emotion recognition and cause analysis in conversations.

Recent related research includes 'Emotion Cause Extraction in Social Media Using Dependency-Based Convolutional Neural Networks' and 'A Multi-Task Approach for Emotion Cause Analysis in Conversations'.