English

Transfer Learning for Speech Recognition on a Budget

Machine Learning 2017-06-02 v1 Computation and Language Neural and Evolutionary Computing Machine Learning

Abstract

End-to-end training of automated speech recognition (ASR) systems requires massive data and compute resources. We explore transfer learning based on model adaptation as an approach for training ASR models under constrained GPU memory, throughput and training data. We conduct several systematic experiments adapting a Wav2Letter convolutional neural network originally trained for English ASR to the German language. We show that this technique allows faster training on consumer-grade resources while requiring less training data in order to achieve the same accuracy, thereby lowering the cost of training ASR models in other languages. Model introspection revealed that small adaptations to the network's weights were sufficient for good performance, especially for inner layers.

Keywords

Cite

@article{arxiv.1706.00290,
  title  = {Transfer Learning for Speech Recognition on a Budget},
  author = {Julius Kunze and Louis Kirsch and Ilia Kurenkov and Andreas Krug and Jens Johannsmeier and Sebastian Stober},
  journal= {arXiv preprint arXiv:1706.00290},
  year   = {2017}
}

Comments

Accepted for 2nd ACL Workshop on Representation Learning for NLP

R2 v1 2026-06-22T20:06:12.372Z