Collaborating Authors


A Double Residual Compression Algorithm for Efficient Distributed Learning Machine Learning

Large-scale machine learning models are often trained by parallel stochastic gradient descent algorithms. However, the communication cost of gradient aggregation and model synchronization between the master and worker nodes becomes the major obstacle for efficient learning as the number of workers and the dimension of the model increase. In this paper, we propose DORE, a DOuble REsidual compression stochastic gradient descent algorithm, to reduce over $95\%$ of the overall communication such that the obstacle can be immensely mitigated. Our theoretical analyses demonstrate that the proposed strategy has superior convergence properties for both strongly convex and nonconvex objective functions. The experimental results validate that DORE achieves the best communication efficiency while maintaining similar model accuracy and convergence speed in comparison with start-of-the-art baselines.

How Apple will stop companies abusing facial recognition on new iPhone X

Boston Herald

When Apple's new iPhone X arrives next month, its Face ID technology will introduce a new era of convenience--but also new risks of broad face-based surveillance by corporations and governments. Apple's strong record on privacy means it's likely to deploy the facial recognition tool responsibly, but that doesn't account for third-party companies that plan to integrate Face ID into their apps. Such companies could seek to assemble their own databases of faces and, in the worst case scenario, use a facial database to identify consumers online and in the streets for ad purposes. Apple has yet to disclose full details of how Face ID will operate, though a source familiar with the tool says there is a plan to prevent app makers from violating user privacy. Meanwhile, outside of a single state law, consumers will have little recourse if companies begin to collect images of their face without consent.