通过逐类微调缓解语音情感识别中的灾难性遗忘问题。
SeQuiFi: Mitigating Catastrophic Forgetting in Speech Emotion Recognition with Sequential Class-Finetuning
- 按情绪类别逐个增量微调模型，避免遗忘旧类
- 在多个数据集上准确率和F1分数显著优于现有方法
- 适合多语言、持续学习场景下的语音情感识别
本文提出SeQuiFi，一种用于缓解语音情感识别中灾难性遗忘的新方法。该方法采用顺序类别微调策略，每次仅对一个情绪类别进行增量微调，从而有效保留并增强各类别的识别能力。尽管已有多种先进方法（如正则化、记忆存储、权重平均等）尝试解决此问题，但在多样性和多语言数据集上仍具挑战。大量实验表明，SeQuiFi在CREMA-D、RAVDESS、Emo-DB、MESD和SHEMO等多个基准数据集上，无论准确率还是F1分数均显著优于原始微调及当前最优持续学习技术。
原文摘要 · Abstract (English)
In this work, we introduce SeQuiFi, a novel approach for mitigating catastrophic forgetting (CF) in speech emotion recognition (SER). SeQuiFi adopts a sequential class-finetuning strategy, where the model is fine-tuned incrementally on one emotion class at a time, preserving and enhancing retention for each class. While various state-of-the-art (SOTA) methods, such as regularization-based, memory-based, and weight-averaging techniques, have been proposed to address CF, it still remains a challenge, particularly with diverse and multilingual datasets. Through extensive experiments, we demonstrate that SeQuiFi significantly outperforms both vanilla fine-tuning and SOTA continual learning techniques in terms of accuracy and F1 scores on multiple benchmark SER datasets, including CREMA-D, RAVDESS, Emo-DB, MESD, and SHEMO, covering different languages.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。