Research map / Optimization / Distributed and compressed
Robust Collaborative Learning with Linear Gradient Overhead
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use MoNNA with local momentum and nearest-neighbor averaging instead of the compared robust collaborative-learning methods | Collaborative learning with faulty nodes · CIFAR-10, FOE, ALIE, LF, and SF attacks | learning accuracy | Helped |
| Use MoNNA instead of SCC | Collaborative learning · CIFAR-10, FOE and SF attacks | learning accuracy: 10% | Helped |
| Use local momentum rather than no momentum (ablation) | MoNNA with (α, λ)-reduction mixing · Non-convex optimization with faulty nodes | asymptotic error dependence on local variance: (f/n) σ^2 -> vanishes at the rate of √(1/T) | Helped |
| Remove momentum (β = 0) (ablation) | MoNNA · Non-convex optimization with faulty nodes | asymptotic error: O(√(σ^2/(nT)) + λ(σ^2 + ζ^2)/(1-α)^2) | Hurt |
| Increase coordination rounds to improve tolerance from n ≥ 11f to n > 5f (ablation) | MoNNA · Non-convex losses | gradient complexity: O(((1/n) + (f/n)) 1/ε^2) -> O((1 + f)^2/(n ε^2)) | Hurt |
Closest papers by meaning
- Robust Distributed Gradient Aggregation Using Projections onto Gradient Manifolds (AAAI 2024 · similarity 0.85)
- Fault Tolerant ML: Efficient Meta-Aggregation and Synchronous Training (ICML 2024 · similarity 0.85)
- From Inexact Gradients to Byzantine Robustness: Acceleration and Optimization under Similarity (NeurIPS 2026 · similarity 0.85)
- A Quadratic Synchronization Rule for Distributed Deep Learning (ICLR 2024 · similarity 0.85)
- A Kernel Perspective on Distillation-based Collaborative Learning (NeurIPS 2024 · similarity 0.84)
- RESIST: Resilient Decentralized Learning Using Consensus Gradient Descent (NeurIPS 2026 · similarity 0.84)
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.