تقرير
Error Feedback Fixes SignSGD and other Gradient Compression Schemes
العنوان: | Error Feedback Fixes SignSGD and other Gradient Compression Schemes |
---|---|
المؤلفون: | Karimireddy, Sai Praneeth, Rebjock, Quentin, Stich, Sebastian U., Jaggi, Martin |
سنة النشر: | 2019 |
المجموعة: | Computer Science Mathematics Statistics |
مصطلحات موضوعية: | Computer Science - Machine Learning, Mathematics - Optimization and Control, Statistics - Machine Learning, I.2.6, I.5.1 |
الوصف: | Sign-based algorithms (e.g. signSGD) have been proposed as a biased gradient compression technique to alleviate the communication bottleneck in training large neural networks across multiple workers. We show simple convex counter-examples where signSGD does not converge to the optimum. Further, even when it does converge, signSGD may generalize poorly when compared with SGD. These issues arise because of the biased nature of the sign compression operator. We then show that using error-feedback, i.e. incorporating the error made by the compression operator into the next step, overcomes these issues. We prove that our algorithm EF-SGD with arbitrary compression operator achieves the same rate of convergence as SGD without any additional assumptions. Thus EF-SGD achieves gradient compression for free. Our experiments thoroughly substantiate the theory and show that error-feedback improves both convergence and generalization. Code can be found at \url{https://github.com/epfml/error-feedback-SGD}. Comment: ICML 2019 (long talk) |
نوع الوثيقة: | Working Paper |
URL الوصول: | http://arxiv.org/abs/1901.09847 |
رقم الأكسشن: | edsarx.1901.09847 |
قاعدة البيانات: | arXiv |
الوصف غير متاح. |