2018

Communication-Efficient On-Device Machine Learning: Federated Distillation and Augmentation under Non-IID Private Data

Jeong, Eunjeong, Oh, Seungeun, Kim, Hyesung et al.

Understand

On-device machine learning (ML) enables the training process to exploit a massive amount of user-generated private data samples.

  • To enjoy this benefit, inter-device communication overhead should be minimized.
  • With this end, we propose federated distillation (FD), a distributed model training algorithm whose communication payload size is much smaller than a benchmark scheme, federated learning (FL), particularly when the model size is large.
  • Moreover, user-generated data samples are likely to become non-IID across devices, which commonly degrades the performance compared to the case with an IID dataset.

Built on

  • Distilling the Knowledge in a Neural Network

    G. E. Hinton, O. Vinyals, and J. Dean · 2014

    Earlier work this paper cites.

  • Tractable Resource Management with Uplink Decoupled Millimeter-Wave Overlay in Ultra-Dense Cellular Networks

    J. Park, S.-L. Kim, and J. Zander · 2016

    Earlier work this paper cites.

Similar

Then

  • Differentially Private Federated Learning: A Client Level Perspective Differentially Private Federated Learning: A Client Level Perspective

    R. C. Geyer, T. Klein, and M. Nabi · 2017

    Later among the works it cites.

  • Smartphones Will Get Even Smarter With On-Device Machine Learning. IEEE Spectrum. Mar. 2018 [Online, Accessed: 2018-11-23]

    M. Bennis · 2018

    Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…