Fine-Grained Emotion Recognition via In-Context Learning

TL;DR

This work reframes fine-grained emotion recognition in in-context learning as a decision-making problem anchored in emotion prototypes. It introduces Emotion In-Context Learning (EICL), which retrieves emotionally similar examples, applies a dynamic soft-label strategy, and uses a two-stage exclusion mechanism to improve both reasoning and final decisions. By validating through multiple datasets and LLMs, the study shows that EICL consistently outperforms standard ICL and zero-shot baselines, particularly when emotional perception varies across models. The results underscore the importance of aligning query representations with emotional prototypes and provide actionable components to enhance practical emotion-enabled systems.

Abstract

Fine-grained emotion recognition aims to identify the emotional type in queries through reasoning and decision-making processes, playing a crucial role in various systems. Recent methods use In-Context Learning (ICL), enhancing the representation of queries in the reasoning process through semantically similar examples, while further improving emotion recognition by explaining the reasoning mechanisms. However, these methods enhance the reasoning process but overlook the decision-making process. This paper investigates decision-making in fine-grained emotion recognition through prototype theory. We show that ICL relies on similarity matching between query representations and emotional prototypes within the model, where emotion-accurate representations are critical. However, semantically similar examples often introduce emotional discrepancies, hindering accurate representations and causing errors. To address this, we propose Emotion In-Context Learning (EICL), which introduces emotionally similar examples and uses a dynamic soft-label strategy to improve query representations in the emotion reasoning process. A two-stage exclusion strategy is then employed to assess similarity from multiple angles, further optimizing the decision-making process. Extensive experiments show that EICL significantly outperforms ICL on multiple datasets.

Figures (10)

• Figure 1: The reasoning and decision-making mechanism of in-context learning (ICL).
• Figure 2: Heatmap of category representations.
• Figure 3: Emotion probability as similarity decreases.
• Figure 4: Overview of EICL. It retrieves emotionally similar examples and uses a dynamic soft-label strategy to accurately depict their emotions, enhancing query representations in the emotion reasoning process. A two-stage exclusion strategy is then applied, prioritizing primary emotion candidates to avoid decision errors from relying solely on similarity, ensuring accurate decision-making.
• Figure 5: Ablation Results for EICL on EDOS and ED Datasets.