In-Context Examples Matter: Improving Emotion Recognition in Conversation with Instruction Tuning

TL;DR

This paper introduces InitERC, a one-stage in-context instruction tuning framework for emotion recognition in conversation, which effectively aligns speaker, context, and emotion cues, outperforming multi-stage methods.

Contribution

Proposes InitERC, a novel one-stage in-context instruction tuning approach that improves emotion recognition by better aligning speaker, context, and emotion information in conversations.

Findings

InitERC significantly outperforms existing baselines.

In-context example selection impacts model performance.

Number and order of examples influence emotion recognition accuracy.

Abstract

Emotion recognition in conversation (ERC) aims to identify the emotion of each utterance in a conversation, playing a vital role in empathetic artificial intelligence. With the growing of large language models (LLMs), instruction tuning has emerged as a critical paradigm for ERC. Existing studies mainly focus on multi-stage instruction tuning, which first endows LLMs with speaker characteristics, and then conducts context-aware instruction tuning to comprehend emotional states. However, these methods inherently constrains the capacity to jointly capture the dynamic interaction between speaker characteristics and conversational context, resulting in weak alignment among speaker identity, contextual cues, and emotion states within a unified framework. In this paper, we propose InitERC, a simple yet effective one-stage in-context instruction tuning framework for ERC. InitERC adapts LLMs to learn speaker-context-emotion alignment from context examples via in-context instruction tuning. Specifically, InitERC comprises four components, i.e., demonstration pool construction, in-context example selection, prompt template design, and in-context instruction tuning. To explore the impact of in-context examples, we conduct a comprehensive study on three key factors: retrieval strategy, example ordering, and the number of examples. Extensive experiments on three widely used datasets demonstrate that our proposed InitERC achieves substantial improvements over the state-of-the-art baselines.