Saved in:
Bibliographic Details
Main Authors: Shi, Hao, Fujita, Yusuke, Koshkin, Roman, Zhao, Mengjie, Gao, Yuan, Liu, Lianbo, Sudo, Yui
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.10587
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918383217278976
author Shi, Hao
Fujita, Yusuke
Koshkin, Roman
Zhao, Mengjie
Gao, Yuan
Liu, Lianbo
Sudo, Yui
author_facet Shi, Hao
Fujita, Yusuke
Koshkin, Roman
Zhao, Mengjie
Gao, Yuan
Liu, Lianbo
Sudo, Yui
contents Large language models (LLMs) provide strong semantic priors that can improve multi-talker automatic speech recognition (MT-ASR), but using an LLM as an autoregressive decoder is computationally expensive and remains fragile under heavy overlap. In this paper, we propose an encoder-only MT-ASR framework that adapts an LLM to multi-talker conditioning and distills its semantic guidance into the encoder during training, while retaining fast CTC-style decoding at inference. Our model employs a post-encoder separator with serialized CTC to produce talker-ordered transcripts, and leverages an adapted LLM-based SOT objective as a multi-talker-aware teacher signal to explicitly regularize mixed-speech representations. To further support variable numbers of talkers, we introduce a Talker-Count Head that predicts the talker count and dynamically selects the appropriate decoding branch. Experiments on LibriMix show that the proposed encoder-only model achieves comparable performance to LLM-based systems in the two-talker condition, while delivering significant improvements in the three-talker condition with significant small RTF.
format Preprint
id arxiv_https___arxiv_org_abs_2603_10587
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Distilling LLM Semantic Priors into Encoder-Only Multi-Talker ASR with Talker-Count Routing
Shi, Hao
Fujita, Yusuke
Koshkin, Roman
Zhao, Mengjie
Gao, Yuan
Liu, Lianbo
Sudo, Yui
Sound
Large language models (LLMs) provide strong semantic priors that can improve multi-talker automatic speech recognition (MT-ASR), but using an LLM as an autoregressive decoder is computationally expensive and remains fragile under heavy overlap. In this paper, we propose an encoder-only MT-ASR framework that adapts an LLM to multi-talker conditioning and distills its semantic guidance into the encoder during training, while retaining fast CTC-style decoding at inference. Our model employs a post-encoder separator with serialized CTC to produce talker-ordered transcripts, and leverages an adapted LLM-based SOT objective as a multi-talker-aware teacher signal to explicitly regularize mixed-speech representations. To further support variable numbers of talkers, we introduce a Talker-Count Head that predicts the talker count and dynamically selects the appropriate decoding branch. Experiments on LibriMix show that the proposed encoder-only model achieves comparable performance to LLM-based systems in the two-talker condition, while delivering significant improvements in the three-talker condition with significant small RTF.
title Distilling LLM Semantic Priors into Encoder-Only Multi-Talker ASR with Talker-Count Routing
topic Sound
url https://arxiv.org/abs/2603.10587