Back to Research papers
Research paper index

AMRD: Adaptive Multi-Teacher Relational Distillation for Lightweight Speech Emotion Recognition

Yuqi Li, Yi-Cheng Lin, Xianglong Wang, Kuo Yang, Xiaoqin Feng, Yixuan Wang, Huiran Duan, Yingli Tian

arXiv:2607.25289Published July 28, 20260 citations
  • cs.LG

Abstract

On-device speech emotion recognition (SER) is critical for real-time applications, yet large self-supervised models that excel at SER are too costly for edge devices. Multi-teacher knowledge distillation can compress them into a lightweight student, but two challenges remain: teacher reliability varies across batches, and logit-level distillation ignores inter-sample relational structure. We propose Adaptive Multi-teacher Relational Distillation (AMRD) to address both. A one-class SVM on each teacher's logit similarity matrix assigns per-batch weights favoring more coherent teachers. A relational distillation loss aligns teacher and student similarity matrices, capturing structure that logit matching misses. On IEMOCAP and CREMA-D datasets across four student architectures, AMRD outperforms single-teacher distillation baselines in most settings, and ablations confirm both components yield complementary gains.

Read the original paper

This page indexes public paper metadata. The manuscript remains with its original publisher and authors.