Research map / Optimization / Distributed and compressed
Escaping Saddle Points with Compressed SGD
Dmitrii Avdiukhin, Grigory Yaroslavtsev
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add artificial noise to the SGD updates (ablation) | Compressed SGD · Deep MNIST autoencoder | saddle-point escape rate | Improved |
| Communicate 10% or 1% of gradient coordinates rather than the full gradient (ablation) | Distributed SGD with RandomK · CIFAR-10, ResNet34 | convergence rate | No clear effect |
| Use RandomK instead of TopK at 0.1% of coordinates (ablation) | Compressed SGD · Deep MNIST autoencoder | saddle-point escape: 0.1% | Mixed |
Closest papers by meaning
- Scalable Distributed Stochastic Optimization via Bidirectional Compression: Beyond Pessimistic Limits (NeurIPS 2026 · similarity 0.91)
- Error Compensated Distributed SGD Can Be Accelerated (NeurIPS 2021 · similarity 0.91)
- Communication-efficient SGD: From Local SGD to One-Shot Averaging (NeurIPS 2021 · similarity 0.91)
- On Biased Compression for Distributed Learning (JMLR 2023 · similarity 0.90)
- STL-SGD: Speeding Up Local SGD with Stagewise Communication Period (AAAI 2021 · similarity 0.90)
- Local SGD Converges Fast and Communicates Little (ICLR 2019 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
