Communication-Efficient Distributed SGD with Error-Feedback, Revisited
We show that the convergence proof of a recent algorithm called dist-EF-SGD for distributed stochastic gradient descent with communication efficiency using error-feedback of Zheng et al., Communication-efficient distributed blockwise momentum SGD with error-feedback, in Advances in Neural Informatio...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2021-04-01
|
| Series: | International Journal of Computational Intelligence Systems |
| Subjects: | |
| Online Access: | https://www.atlantis-press.com/article/125955624/view |
| _version_ | 1828342347691196416 |
|---|---|
| author | Tran Thi Phuong Le Trieu Phong |
| author_facet | Tran Thi Phuong Le Trieu Phong |
| author_sort | Tran Thi Phuong |
| collection | DOAJ |
| description | We show that the convergence proof of a recent algorithm called dist-EF-SGD for distributed stochastic gradient descent with communication efficiency using error-feedback of Zheng et al., Communication-efficient distributed blockwise momentum SGD with error-feedback, in Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019 (NeurIPS 2019), 2019, pp. 11446–11456, is problematic mathematically. Concretely, the original error bound for arbitrary sequences of learning rate is unfortunately incorrect, leading to an invalidated upper bound in the convergence theorem for the algorithm. As evidences, we explicitly provide several counter-examples, for both convex and nonconvex cases, to show the incorrectness of the error bound. We fix the issue by providing a new error bound and its corresponding proof, leading to a new convergence theorem for the dist-EF-SGD algorithm, and therefore recovering its mathematical analysis. |
| first_indexed | 2024-04-13T23:28:32Z |
| format | Article |
| id | doaj.art-05d419a6740940bd94d1b97e64b6efd2 |
| institution | Directory Open Access Journal |
| issn | 1875-6883 |
| language | English |
| last_indexed | 2024-04-13T23:28:32Z |
| publishDate | 2021-04-01 |
| publisher | Springer |
| record_format | Article |
| series | International Journal of Computational Intelligence Systems |
| spelling | doaj.art-05d419a6740940bd94d1b97e64b6efd22022-12-22T02:24:59ZengSpringerInternational Journal of Computational Intelligence Systems1875-68832021-04-0114110.2991/ijcis.d.210412.001Communication-Efficient Distributed SGD with Error-Feedback, RevisitedTran Thi PhuongLe Trieu PhongWe show that the convergence proof of a recent algorithm called dist-EF-SGD for distributed stochastic gradient descent with communication efficiency using error-feedback of Zheng et al., Communication-efficient distributed blockwise momentum SGD with error-feedback, in Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019 (NeurIPS 2019), 2019, pp. 11446–11456, is problematic mathematically. Concretely, the original error bound for arbitrary sequences of learning rate is unfortunately incorrect, leading to an invalidated upper bound in the convergence theorem for the algorithm. As evidences, we explicitly provide several counter-examples, for both convex and nonconvex cases, to show the incorrectness of the error bound. We fix the issue by providing a new error bound and its corresponding proof, leading to a new convergence theorem for the dist-EF-SGD algorithm, and therefore recovering its mathematical analysis.https://www.atlantis-press.com/article/125955624/viewOptimizerDistributed learningSGDError-feedbackDeep neural networks |
| spellingShingle | Tran Thi Phuong Le Trieu Phong Communication-Efficient Distributed SGD with Error-Feedback, Revisited International Journal of Computational Intelligence Systems Optimizer Distributed learning SGD Error-feedback Deep neural networks |
| title | Communication-Efficient Distributed SGD with Error-Feedback, Revisited |
| title_full | Communication-Efficient Distributed SGD with Error-Feedback, Revisited |
| title_fullStr | Communication-Efficient Distributed SGD with Error-Feedback, Revisited |
| title_full_unstemmed | Communication-Efficient Distributed SGD with Error-Feedback, Revisited |
| title_short | Communication-Efficient Distributed SGD with Error-Feedback, Revisited |
| title_sort | communication efficient distributed sgd with error feedback revisited |
| topic | Optimizer Distributed learning SGD Error-feedback Deep neural networks |
| url | https://www.atlantis-press.com/article/125955624/view |
| work_keys_str_mv | AT tranthiphuong communicationefficientdistributedsgdwitherrorfeedbackrevisited AT letrieuphong communicationefficientdistributedsgdwitherrorfeedbackrevisited |