English

Federated Self-Learning with Weak Supervision for Speech Recognition

Audio and Speech Processing 2023-06-22 v1 Sound

Abstract

Automatic speech recognition (ASR) models with low-footprint are increasingly being deployed on edge devices for conversational agents, which enhances privacy. We study the problem of federated continual incremental learning for recurrent neural network-transducer (RNN-T) ASR models in the privacy-enhancing scheme of learning on-device, without access to ground truth human transcripts or machine transcriptions from a stronger ASR model. In particular, we study the performance of a self-learning based scheme, with a paired teacher model updated through an exponential moving average of ASR models. Further, we propose using possibly noisy weak-supervision signals such as feedback scores and natural language understanding semantics determined from user behavior across multiple turns in a session of interactions with the conversational agent. These signals are leveraged in a multi-task policy-gradient training approach to improve the performance of self-learning for ASR. Finally, we show how catastrophic forgetting can be mitigated by combining on-device learning with a memory-replay approach using selected historical datasets. These innovations allow for 10% relative improvement in WER on new use cases with minimal degradation on other test sets in the absence of strong-supervision signals such as ground-truth transcriptions.

Keywords

Cite

@article{arxiv.2306.12015,
  title  = {Federated Self-Learning with Weak Supervision for Speech Recognition},
  author = {Milind Rao and Gopinath Chennupati and Gautam Tiwari and Anit Kumar Sahu and Anirudh Raju and Ariya Rastrow and Jasha Droppo},
  journal= {arXiv preprint arXiv:2306.12015},
  year   = {2023}
}

Comments

Proceedings of ICASSP 2023

R2 v1 2026-06-28T11:10:22.144Z