Emotion-Anchored Contrastive Learning Framework for Emotion Recognition in Conversation

TL;DR

This work tackles emotion recognition in conversation (ERC) by addressing the difficulty of distinguishing similar emotions in context. It introduces Emotion-Anchored Contrastive Learning (EACL), which uses label-informed emotion anchors in a two-stage framework: Stage One learns discriminative utterance representations guided by anchors via a specialized contrastive objective and cross-entropy loss, while Stage Two adapts anchor positions to better align with learned representations. Across three benchmark datasets, EACL achieves state-of-the-art results and demonstrates enhanced separability for similar emotions, with notable improvements such as larger gains between pairs like excited vs happy. The approach is robust across multiple language models and suggests practical impact for dialogue systems requiring fine-grained emotional understanding.

Abstract

Emotion Recognition in Conversation (ERC) involves detecting the underlying emotion behind each utterance within a conversation. Effectively generating representations for utterances remains a significant challenge in this task. Recent works propose various models to address this issue, but they still struggle with differentiating similar emotions such as excitement and happiness. To alleviate this problem, We propose an Emotion-Anchored Contrastive Learning (EACL) framework that can generate more distinguishable utterance representations for similar emotions. To achieve this, we utilize label encodings as anchors to guide the learning of utterance representations and design an auxiliary loss to ensure the effective separation of anchors for similar emotions. Moreover, an additional adaptation process is proposed to adapt anchors to serve as effective classifiers to improve classification performance. Across extensive experiments, our proposed EACL achieves state-of-the-art emotion recognition performance and exhibits superior performance on similar emotions. Our code is available at https://github.com/Yu-Fangxu/EACL.

Figures (7)

• Figure 1: An example of a conversation in the IEMOCAP dataset.
• Figure 2: Normalized confusion matrix of SPCL on the IEMOCAP dataset. The rows and columns represent the actual classes and predictions made by the model respectively. The cross-point ($i$, $j$) means the percentage of emotion $i$ predicted to be emotion $j$. Except for the diagonal, the bigger values and deeper color mean these emotions are easily misclassified.
• Figure 3: Overview of our proposed framework. Left side introduces representation learning, which is composed of utterance representation and emotion anchor learning. Right side describes the process of adapting emotion anchors to the optimal positions for classification.
• Figure 4: The cosine similarity of pair-wise emotions. Figure (a) and (b) depicts cosine similarity between emotion anchors before and after training with EACL. (c) and (d) depicts the angle degree between emotion anchors before and after training with EACL respectively.
• Figure 5: The t-SNE visualization of emotion anchors. Circles represent the position of emotion anchors before training and stars are the positions after training.