Corpus ID: 218763327

Training Keyword Spotting Models on Non-IID Data with Federated Learning

@article{Hard2020TrainingKS,
  title={Training Keyword Spotting Models on Non-IID Data with Federated Learning},
  author={Andrew Hard and Kurt Partridge and Cameron Nguyen and Niranjan Subrahmanya and Aishanee Shah and Pai Zhu and Ignacio Lopez-Moreno and Rajiv Mathews},
  journal={ArXiv},
  year={2020},
  volume={abs/2005.10406}
}
  • Andrew Hard, Kurt Partridge, +5 authors Rajiv Mathews
  • Published 2020
  • Computer Science, Engineering
  • ArXiv
  • We demonstrate that a production-quality keyword-spotting model can be trained on-device using federated learning and achieve comparable false accept and false reject rates to a centrally-trained model. To overcome the algorithmic constraints associated with fitting on-device data (which are inherently non-independent and identically distributed), we conduct thorough empirical studies of optimization algorithms and hyperparameter configurations using large-scale federated simulations. To… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Citations

    Publications citing this paper.

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 44 REFERENCES

    Federated Learning for Keyword Spotting

    VIEW 2 EXCERPTS

    Federated Learning with Non-IID Data

    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL

    Adaptive Federated Optimization

    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL

    Advances and Open Problems in Federated Learning

    VIEW 1 EXCERPT

    End-to-End Streaming Keyword Spotting

    VIEW 3 EXCERPTS

    Small-footprint keyword spotting using deep neural networks

    VIEW 1 EXCERPT