# Pseudo2Real: Task Arithmetic for Pseudo-Label Correction in Automatic Speech Recognition

arXiv: https://arxiv.org/abs/2510.08047 · HTML: https://mlsanity.com/paper/2510.08047/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 44 references that are in the map.
- First posted on arXiv 9 Oct 2025.

## Builds on (most cited first, 24 of 24)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Linear Mode Connectivity and the Lottery Ticket Hypothesis](https://mlsanity.com/paper/1912.05671/) (2019)
- [ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification](https://mlsanity.com/paper/2005.07143/) (2020)
- [What is being transferred in transfer learning?](https://mlsanity.com/paper/2008.11687/) (2020)
- [Improved Noisy Student Training for Automatic Speech Recognition](https://mlsanity.com/paper/2005.09629/) (2020)
- [When is Task Vector Provably Effective for Model Editing? A Generalization Analysis of Nonlinear Transformers](https://mlsanity.com/paper/2504.10957/) (2025)
- [AfriSpeech-200: Pan-African Accented Speech Dataset for Clinical and General Domain ASR](https://mlsanity.com/paper/2310.00274/) (2023)
- [Codec-SUPERB: An In-Depth Analysis of Sound Codec Models](https://mlsanity.com/paper/2402.13071/) (2024)
- [Building a Taiwanese Mandarin Spoken Language Model: A First Attempt](https://mlsanity.com/paper/2411.07111/) (2024)
- [Kaizen: Continuously improving teacher using Exponential Moving Average for semi-supervised speech recognition](https://mlsanity.com/paper/2106.07759/) (2021)
- [Semi-Supervised Speech Recognition via Graph-based Temporal Classification](https://mlsanity.com/paper/2010.15653/) (2020)
- [Continuous Soft Pseudo-Labeling in ASR](https://mlsanity.com/paper/2211.06007/) (2022)
- [Low-Rank and Sparse Model Merging for Multi-Lingual Speech Recognition and Translation](https://mlsanity.com/paper/2502.17380/) (2025)
- [Distilling a speech and music encoder with task arithmetic](https://mlsanity.com/paper/2505.13270/) (2025)
- [Better Pseudo-labeling with Multi-ASR Fusion and Error Correction by SpeechLLM](https://mlsanity.com/paper/2506.11089/) (2025)
- [Unsupervised domain adaptation for speech recognition with unsupervised error correction](https://mlsanity.com/paper/2209.12043/) (2022)
- [Federated Self-Learning with Weak Supervision for Speech Recognition](https://mlsanity.com/paper/2306.12015/) (2023)
- [Mitigating Subgroup Disparities in Multi-Label Speech Emotion Recognition: A Pseudo-Labeling and Unsupervised Learning Approach](https://mlsanity.com/paper/2505.14449/) (2025)
- [ToxicTone: A Mandarin Audio Dataset Annotated for Toxicity and Toxic Utterance Tonality](https://mlsanity.com/paper/2505.15773/) (2025)
- [Progressive unsupervised domain adaptation for ASR using ensemble models and multi-stage training](https://mlsanity.com/paper/2402.04805/) (2024)
- [Conformer-1: Robust ASR via Large-Scale Semisupervised Bootstrapping](https://mlsanity.com/paper/2404.07341/) (2024)
- [Sotto Voce: Federated Speech Recognition with Differential Privacy Guarantees](https://mlsanity.com/paper/2207.07816/) (2022)
- [Speech-FT: Merging Pre-trained And Fine-Tuned Speech Representation Models For Cross-Task Generalization](https://mlsanity.com/paper/2502.12672/) (2025)
- [A correlation-permutation approach for speech-music encoders model merging](https://mlsanity.com/paper/2506.11403/) (2025)

## Built on by (most cited first, 1 of 1)

- [ReHear: Iterative Pseudo-Label Refinement for Semi-Supervised Speech Recognition via Audio Large Language Models](https://mlsanity.com/paper/2602.18721/) (2026)

## Abstract (arXiv)

Robust ASR under domain shift is crucial because real-world systems encounter unseen accents and domains with limited labeled data. Although pseudo-labeling offers a practical workaround, it often introduces systematic, accent-specific errors that filtering fails to fix. We ask: How can we correct these recurring biases without target ground truth? We propose a simple parameter-space correction: in a source domain containing both real and pseudo-labeled data, two ASR models are fine-tuned from the same initialization, one on ground-truth labels and the other on pseudo-labels, and their weight difference forms a correction vector that captures pseudo-label biases. When applied to a pseudo-labeled target model, this vector enhances recognition, achieving up to a 35% relative Word Error Rate (WER) reduction on AfriSpeech-200 across ten African accents with the Whisper tiny model.