# Self-Taught Recognizer: Toward Unsupervised Adaptation for Speech Foundation Models

arXiv: https://arxiv.org/abs/2405.14161 · HTML: https://mlsanity.com/paper/2405.14161/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 50 of its 95 references that are in the map.
- First posted on arXiv 23 May 2024.

## Builds on (most cited first, 30 of 50)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?](https://mlsanity.com/paper/1703.04977/) (2017)
- [Unsupervised Domain Adaptation by Backpropagation](https://mlsanity.com/paper/1409.7495/) (2014)
- [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://mlsanity.com/paper/2110.13900/) (2021)
- [Common Voice: A Massively-Multilingual Speech Corpus](https://mlsanity.com/paper/1912.06670/) (2019)
- [Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation](https://mlsanity.com/paper/2002.08546/) (2020)
- [OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation](https://mlsanity.com/paper/2311.17911/) (2023)
- [FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech](https://mlsanity.com/paper/2205.12446/) (2022)
- [Analyzing the Structure of Attention in a Transformer Language Model](https://mlsanity.com/paper/1906.04284/) (2019)
- [Lip Reading Sentences in the Wild](https://mlsanity.com/paper/1611.05358/) (2016)
- [TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptation](https://mlsanity.com/paper/1805.04699/) (2018)
- [Universal Source-Free Domain Adaptation](https://mlsanity.com/paper/2004.04393/) (2020)
- [Unsupervised Quality Estimation for Neural Machine Translation](https://mlsanity.com/paper/2005.10608/) (2020)
- [State-of-the-art Speech Recognition With Sequence-to-Sequence Models](https://mlsanity.com/paper/1712.01769/) (2017)
- [Seamless: Multilingual Expressive and Streaming Speech Translation](https://mlsanity.com/paper/2312.05187/) (2023)
- [Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation](https://mlsanity.com/paper/2110.04202/) (2021)
- [Recall and Learn: Fine-tuning Deep Pretrained Language Models with Less Forgetting](https://mlsanity.com/paper/2004.12651/) (2020)
- [Evaluating model calibration in classification](https://mlsanity.com/paper/1902.06977/) (2019)
- [Model Adaptation: Historical Contrastive Learning for Unsupervised Domain Adaptation without Source Data](https://mlsanity.com/paper/2110.03374/) (2021)
- [CHiME-6 Challenge:Tackling Multispeaker Speech Recognition for Unsegmented Recordings](https://mlsanity.com/paper/2004.09249/) (2020)
- [Source-Free Domain Adaptation via Distribution Estimation](https://mlsanity.com/paper/2204.11257/) (2022)
- [Source-Free Unsupervised Domain Adaptation: A Survey](https://mlsanity.com/paper/2301.00265/) (2022)
- [A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Data](https://mlsanity.com/paper/2012.05400/) (2020)
- [OWSM v3.1: Better and Faster Open Whisper-Style Speech Models based on E-Branchformer](https://mlsanity.com/paper/2401.16658/) (2024)
- [Generative Speech Recognition Error Correction with Large Language Models and Task-Activating Prompting](https://mlsanity.com/paper/2309.15649/) (2023)
- [Source-Free Domain Adaptive Fundus Image Segmentation with Denoised Pseudo-Labeling](https://mlsanity.com/paper/2109.09735/) (2021)
- [HyPoradise: An Open Baseline for Generative Speech Recognition with Large Language Models](https://mlsanity.com/paper/2309.15701/) (2023)
- [Unsupervised Domain Adaptation for Speech Recognition via Uncertainty Driven Self-Training](https://mlsanity.com/paper/2011.13439/) (2020)
- [Confidence Estimation for Attention-based Sequence-to-sequence Models for Speech Recognition](https://mlsanity.com/paper/2010.11428/) (2020)

## Built on by (most cited first, 8 of 8)

- [Failing Forward: Improving Generative Error Correction for ASR with Synthetic Data and Retrieval Augmentation](https://mlsanity.com/paper/2410.13198/) (2024)
- [Token-Weighted RNN-T for Learning from Flawed Data](https://mlsanity.com/paper/2406.18108/) (2024)
- [Dual-branch Graph Domain Adaptation for Cross-scenario Multi-modal Emotion Recognition](https://mlsanity.com/paper/2603.26840/) (2026)
- [Customizing Speech Recognition Model with Large Language Model Feedback](https://mlsanity.com/paper/2506.11091/) (2025)
- [Speech Recognition Model Improves Text-to-Speech Synthesis using Fine-Grained Reward](https://mlsanity.com/paper/2511.17555/) (2025)
- [Gumbel-BEARD: Automatic Layer Selection for Self-Supervised Adaptation of Whisper in Low-Resource Domains](https://mlsanity.com/paper/2606.11429/) (2026)
- [On-Policy Self-Distillation for Multi-Dialect ASR: Mastering Dialects, Retaining Mandarin](https://mlsanity.com/paper/2608.11898/) (2026)
- [Pretrained ASR Pseudo-labeling for Noisy Police Audio](https://mlsanity.com/paper/2609.30469/) (2026)

## Abstract (arXiv)

We propose an unsupervised adaptation framework, Self-TAught Recognizer (STAR), which leverages unlabeled data to enhance the robustness of automatic speech recognition (ASR) systems in diverse target domains, such as noise and accents. STAR is developed for prevalent speech foundation models based on Transformer-related architecture with auto-regressive decoding (e.g., Whisper, Canary). Specifically, we propose a novel indicator that empirically integrates step-wise information during decoding to assess the token-level quality of pseudo labels without ground truth, thereby guiding model updates for effective unsupervised adaptation. Experimental results show that STAR achieves an average of 13.5% relative reduction in word error rate across 14 target domains, and it sometimes even approaches the upper-bound performance of supervised adaptation. Surprisingly, we also observe that STAR prevents the adapted model from the common catastrophic forgetting problem without recalling source-domain data. Furthermore, STAR exhibits high data efficiency that only requires less than one-hour unlabeled data, and seamless generality to alternative large speech models and speech translation tasks. Our code aims to open source to the research communities.