# Choosing a PEFT Variant for Per-Patient Dysarthric ASR: A Single-Speaker Case Study on Two ASR Bases

arXiv: https://arxiv.org/abs/2609.02735 · HTML: https://mlsanity.com/paper/2609.02735/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 15 of its 30 references that are in the map.
- First posted on arXiv 2 Sep 2026.

## Builds on (most cited first, 15 of 15)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [DoRA: Weight-Decomposed Low-Rank Adaptation](https://mlsanity.com/paper/2402.09353/) (2024)
- [FedPara: Low-Rank Hadamard Product for Communication-Efficient Federated Learning](https://mlsanity.com/paper/2108.06098/) (2021)
- [VB-LoRA: Extreme Parameter Efficient Fine-Tuning with Vector Banks](https://mlsanity.com/paper/2405.15179/) (2024)
- [Personalizing ASR for Dysarthric and Accented Speech with Limited Data](https://mlsanity.com/paper/1907.13511/) (2019)
- [Personalized Automatic Speech Recognition Trained on Small Disordered Speech Datasets](https://mlsanity.com/paper/2110.04612/) (2021)
- [Parameter-efficient Dysarthric Speech Recognition Using Adapter Fusion and Householder Transformation](https://mlsanity.com/paper/2306.07090/) (2023)
- [Personalized Fine-Tuning with Controllable Synthetic Speech from LLM-Generated Transcripts for Dysarthric Speech Recognition](https://mlsanity.com/paper/2505.12991/) (2025)
- [Perceiver-Prompt: Flexible Speaker Adaptation in Whisper for Chinese Disordered Speech Recognition](https://mlsanity.com/paper/2406.09873/) (2024)
- [Variational Low-Rank Adaptation for Personalized Impaired Speech Recognition](https://mlsanity.com/paper/2509.20397/) (2025)
- [Improved Dysarthric Speech to Text Conversion via TTS Personalization](https://mlsanity.com/paper/2508.06391/) (2025)
- [FiLM-Based Speaker Conditioning of a SpeechLLM for Pathological Speech Recognition](https://mlsanity.com/paper/2606.06211/) (2026)
- [Adapting Foundation ASR Models to Dysarthric Speech: A Case Study](https://mlsanity.com/paper/2606.31722/) (2026)

## Abstract (arXiv)

Per-patient adapters are the preferred production architecture for dysarthric automatic speech recognition (ASR), yet parameter-efficient fine-tuning (PEFT) variants have not been compared in the speaker-dependent, per-patient regime. We present a single-speaker case study comparing seven LoRA-family methods (LoRA, QLoRA, AdaLoRA, DoRA, LoHA, VeRA, VB-LoRA) on two production bases (Whisper-large-v3 with Hungarian fine-tuning, and a multilingual Qwen3-ASR-1.7B checkpoint) for one post-stroke Hungarian male speaker (S1, 409 utterances; severe dysarthria on auditory-perceptual clinical assessment). Attention-projection adapters substantially improve CER on both bases. Across three seeds, a paired bootstrap detects no significant LoRA-DoRA difference (p>0.5; 13.86/13.90 % CER on Whisper, 28.10/28.33 % on Qwen3-ASR), so we adopt the simpler, cheaper LoRA. Real 4-bit (NF4) QLoRA is worse on every seed and both bases (14.56/30.09 % CER) with no memory saving at this scale, and LoHA, VeRA, VB-LoRA and AdaLoRA do not reach the LoRA family, though LoHA still gives an 18.6 % relative CER reduction on Whisper. On the same base, full fine-tuning is more accurate (11.43 % CER), but a 115 MB LoRA that also adapts the feed-forward blocks reaches within 0.66 pp of it at approximately 3.7 % of the per-patient storage. A 6-point enrollment grid shows about 5 min of patient audio captures 45.6 % of the zero-shot-to-30-min CER reduction, with further gains at 10 and 30 min (caveat: one speaker, one language, severe post-stroke dysarthria). Training scripts and recipes will be released, source-available under a research-use licence, on publication.