Research map / Learning theory and algorithms / Prediction and decision losses
Coded Distributed Computing for Inverse Problems
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 |
|---|---|---|---|
| Coded linear inverse (Algorithm 1) instead of replication or uncoded solvers | personalized PageRank on Google Plus graph · 120 workers, k=100 PageRank problems, deadline sweep | average mean-squared error: MSE ratio about 10^4 at large deadlines | Helped |
| Coded linear inverse instead of replication-based solvers | personalized PageRank on Twitter graph · 120 workers, k=100, deadline Tdl up to 2 s | average mean-squared error ratio: ratio grows with deadline; up to 10^4 reported in abstract | Helped |
| Using the extended matrix-inverse decoding of Lee et al. [6] (decoding from only fastest k partial results) instead of Algorithm 1 (ablation) | coded PageRank on Google Plus graph · 120 workers, k=100 | MSE | Hurt |
| Using all partial results including stragglers (weighted decoding) instead of ignoring stragglers as erasures | coded linear inverse · PageRank experiments | MSE | Helped |
| Choice of generator matrix: DFT code vs binary, sparse, Gaussian codes (ablation) | coded linear inverse · Twitter graph PageRank | MSE | Helped |
| Replication with n-k < k extra workers (vs uncoded) | linear inverse solvers · theoretical error exponent as Tdl to infinity | error exponent | No clear effect |
Closest papers by meaning
- Coded Computing for Resilient Distributed Computing: A Learning-Theoretic Framework (NeurIPS 2024 · similarity 0.87)
- Leveraging partial stragglers within gradient coding (NeurIPS 2024 · similarity 0.87)
- Gradient Coding: Avoiding Stragglers in Distributed Learning (ICML 2017 · similarity 0.86)
- Accelerated Methods with Compressed Communications for Distributed Optimization Problems Under Data Similarity (AAAI 2025 · similarity 0.85)
- Approximate Gradient Coding for Distributed Learning with Heterogeneous Stragglers (NeurIPS 2025 · similarity 0.85)
- Rescaled Asynchronous SGD: Optimal Distributed Optimization under Data and System Heterogeneity (NeurIPS 2026 · similarity 0.83)
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.