Research map / Bandits and online learning / Online convex optimization and games
Distributed Primal-Dual Optimization for Online Multi-Task Learning
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 leading singular vectors instead of a full SVD | DROM · Online multi-task learning | time complexity: O(dm^2) to O(dm) | Helped |
| Use primal-dual updates instead of online Frank-Wolfe | DROM · Online multi-task learning | regret order: T^{3/4} -> T^{1/2} | Helped |
| Use DROM instead of D-ADMM | DROM · Spam Email, MHC-I, EachMovie | runtime (sec/iteration) | Helped |
| Use capped Lp loss to downweight large-loss outliers | DROM · Adversarial noisy labels | classification performance | Helped |
| Increase the network sparsity parameter ζ (ablation) | DROM-D · Spam Email, MHC-I, EachMovie | error rate | Hurt |
| Increase the network sparsity parameter ζ (ablation) | DROM-D · Spam Email, MHC-I, EachMovie | F1-measure | Hurt |
| Increase the synchronization interval from 1 to 20 in a fully connected network (ablation) | DROM-D · Spam Email, MHC-I, EachMovie | error rate | No clear effect |
| Increase the synchronization interval from 1 to 20 in a fully connected network (ablation) | DROM-D · Spam Email, MHC-I, EachMovie | F1-measure | No clear effect |
| Synchronize every 20 rounds instead of every round in a rid-connected network (ablation) | DROM-D · Spam Email, MHC-I, EachMovie | error rate | Hurt |
| Synchronize every 20 rounds instead of every round in a ring-connected network (ablation) | DROM-D · Spam Email, MHC-I, EachMovie | F1-measure | Hurt |
| Learn task relatedness jointly rather than train each task locally | DROM · Spam Email, MHC-I, EachMovie | prediction performance | Helped |
Closest papers by meaning
- Convergence of Distributed Adaptive Optimization with Local Updates (ICLR 2025 · similarity 0.86)
- ATA: Adaptive Task Allocation for Efficient Resource Management in Distributed Machine Learning (ICML 2025 · similarity 0.86)
- Online Learning under Budget and ROI Constraints via Weak Adaptivity (ICML 2024 · similarity 0.85)
- Discounted Adaptive Online Learning: Towards Better Regularization (ICML 2024 · similarity 0.85)
- AdaMerging: Adaptive Model Merging for Multi-Task Learning (ICLR 2024 · similarity 0.84)
- Delve into the Applicability of Advanced Optimizers for Multi-Task Learning (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.