Learning to Prefer Reliably: Error-Augmented Emotion Preference Optimization with Calibrated Fusion

2026-08-25

AI Summary

本文提出Error-Augmented Preference Optimization方法，通过增加错误样本和多模型融合解决情绪偏好学习中数据稀疏及模型偏见问题。

Abstract

Emotion preference learning uses pairwise comparisons between candidate descriptions to align multimodal large language models (MLLMs) with human judgments of open-ended emotion descriptions and to train reward models that capture human emotional preferences. However, conventional pairwise supervision is often sparse, typically providing only a single negative description for each positive description, and therefore offers limited coverage of the diverse ways in which an emotion description can be incorrect. In particular, models may be insufficiently exposed to semantically fluent but emotionally inconsistent descriptions. Beyond this data-level limitation, relying on a single MLLM judge introduces a distinct model-level concern: its judgments can be affected by model-specific biases when interpreting fine-grained or ambiguous multimodal emotional cues. To address these limitations, we propose Error-Augmented Preference Optimization (EAPO), a framework for improving the reliability of MLLM-based emotion preference judgment at both the data and model levels. First, we construct an error-augmented dataset by generating multiple controlled and emotion-aware negative descriptions from each preferred description. We then adapt multiple independent MLLM judges to this richer supervision and aggregate their preference margins using margin-calibrated soft fusion, which maps heterogeneous margins to a common scale before aggregation. Experiments on the MER2026-EmoPrefer Challenge dataset and our error-augmented dataset demonstrate that EAPO improves emotion preference prediction and enhances the robustness of MLLM judges when evaluating fluent descriptions that conflict with the video's multimodal emotional evidence. Our code is available at https://github.com/slash1028/EAPO-EmoPrefer.

Problem

• emotion preference learning
• pairwise comparisons
• multimodal large language models
• sparse supervision
• model-specific biases

Innovation

• Error-Augmented Preference Optimization
• emotion preference learning
• margin-calibrated soft fusion
• multimodal large language models