# RobustDistiller: Compressing Universal Speech Representations for Enhanced Environment Robustness

arXiv: https://arxiv.org/abs/2302.09437 · HTML: https://mlsanity.com/paper/2302.09437/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Feb 2023.

## Built on by (most cited first, 5 of 5)

- [DiTSE: High-Fidelity Generative Speech Enhancement via Latent Diffusion Transformers](https://mlsanity.com/paper/2504.09381/) (2025)
- [On the Transferability of Whisper-based Representations for "In-the-Wild" Cross-Task Downstream Speech Applications](https://mlsanity.com/paper/2305.14546/) (2023)
- [VIC-KD: Variance-Invariance-Covariance Knowledge Distillation to Make Keyword Spotting More Robust Against Adversarial Attacks](https://mlsanity.com/paper/2309.12914/) (2023)
- [Alethia: A Foundational Encoder for Voice Deepfakes](https://mlsanity.com/paper/2605.00251/) (2026)
- [Noise robust distillation of self-supervised speech models via correlation metrics](https://mlsanity.com/paper/2312.12153/) (2023)

## Abstract (arXiv)

Self-supervised speech pre-training enables deep neural network models to capture meaningful and disentangled factors from raw waveform signals. The learned universal speech representations can then be used across numerous downstream tasks. These representations, however, are sensitive to distribution shifts caused by environmental factors, such as noise and/or room reverberation. Their large sizes, in turn, make them unfeasible for edge applications. In this work, we propose a knowledge distillation methodology termed RobustDistiller which compresses universal representations while making them more robust against environmental artifacts via a multi-task learning objective. The proposed layer-wise distillation recipe is evaluated on top of three well-established universal representations, as well as with three downstream tasks. Experimental results show the proposed methodology applied on top of the WavLM Base+ teacher model outperforming all other benchmarks across noise types and levels, as well as reverberation times. Oftentimes, the obtained results with the student model (24M parameters) achieved results inline with those of the teacher model (95M).