Extracting generalized and robust representations is a major challenge in emotion recognition in conversations (ERC). To address this, we propose a supervised adversarial contrastive learning (SACL) framework for learning class-spread structured representations. The framework applies contrast-aware adversarial training to generate worst-case samples and uses a joint class-spread contrastive learning objective on both original and adversarial samples. It can effectively utilize label-level feature consistency and retain fine-grained intra-class features. To avoid the negative impact of adversarial perturbations on context-dependent data, we design a contextual adversarial training strategy to learn more diverse features from context and enhance the model's context robustness. We develop a sequence-based method SACL-LSTM under this framework, to learn label-consistent and context-robust emotional features for ERC. Experiments on three datasets demonstrate that SACL-LSTM achieves state-of-the-art performance on ERC. Extended experiments prove the effectiveness of the SACL framework.

翻译：提取通用且鲁棒的表征是对话情感识别（ERC）中的一项主要挑战。为此，我们提出了一种监督式对抗对比学习（SACL）框架，用于学习类别分散的结构化表征。该框架通过对比感知对抗训练生成最坏情况样本，并在原始样本与对抗样本上联合应用类别分散对比学习目标，从而有效利用标签级别的特征一致性，同时保留细粒度的类内特征。为避免对抗扰动对上下文相关数据产生负面影响，我们设计了上下文感知对抗训练策略，从上下文中学习更多样化的特征，并增强模型的上下文鲁棒性。在此框架下，我们开发了一种基于序列的方法SACL-LSTM，用于学习标签一致且上下文鲁棒的情感特征以支持对话情感识别。在三个数据集上的实验表明，SACL-LSTM在对话情感识别任务上达到了最优性能。扩展实验进一步证明了SACL框架的有效性。