Research map / Optimization / Distributed and compressed
Quantized Decentralized Stochastic Learning over Directed Graphs
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 |
|---|---|---|---|
| Quantize the difference x_i - x̂_i (with stochastic low-precision quantizer) in push-sum gossip over directed graphs | Quantized push-sum gossip · Gossip over directed graphs G1, G2 | total communicated bits to reach target error: up to 10x (G1) and 6x (G2) fewer bits | Helped |
| Quantized push-sum gossip versus exact communication, measured per iteration | Quantized push-sum gossip · Gossip over directed graphs G1, G2 | error per iteration | Hurt |
| Quantized decentralized SGD with fewer bits per entry versus exact communication | Quantized decentralized SGD (Algorithm 2) · Least-squares objective, d=256, G1/G2 | loss per iteration and bits to reach loss: 5x fewer bits | Mixed |
| 8-bit quantization in decentralized SGD on MNIST neural network | Quantized decentralized SGD (Algorithm 2) · MNIST training loss | training loss per iteration and total bits: 5x fewer total bits; per-iteration gap is close | Mixed |
Closest papers by meaning
- Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication (ICML 2019 · similarity 0.90)
- Beyond Exponential Graph: Communication-Efficient Topologies for Decentralized Learning via Finite-time Convergence (NeurIPS 2023 · similarity 0.89)
- Robust and Communication-Efficient Collaborative Learning (NeurIPS 2019 · similarity 0.89)
- Enhancing Parallelism in Decentralized Stochastic Convex Optimization (ICML 2025 · similarity 0.88)
- Towards Faster Decentralized Stochastic Optimization with Communication Compression (ICLR 2025 · similarity 0.87)
- Achieving Linear Speedup and Near-Optimal Complexity for Decentralized Optimization over Row-stochastic Networks (ICML 2025 · similarity 0.87)
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.