Research map / Bandits and online learning / Online convex optimization and games
Efficient Projection-Free Online Methods with Stochastic Recursive Gradient
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 ORGFW instead of OSFW or Regularized OFW | stochastic online multiclass logistic regression · MNIST and CIFAR10 | regret | Helped |
| Use ORGFW instead of OFW or OAW | stochastic online multiclass logistic regression · MNIST and CIFAR10 | regret | Hurt |
| Use ORGFW instead of OFW or OAW | stochastic online multiclass logistic regression · MNIST and CIFAR10 | per-iteration running time | Helped |
| Use ORGFW instead of the compared methods | stochastic optimization · MNIST and CIFAR10 multiclass logistic regression | suboptimality versus running time | Helped |
| Use MORGFW instead of Meta-Frank-Wolfe | adversarial online multiclass logistic regression · MNIST and CIFAR10 | regret | Mixed |
| Use MORGFW instead of OFW, OAW, or Regularized OFW | adversarial online multiclass logistic regression · MNIST and CIFAR10 | regret | Helped |
| Use MORGFW instead of Regularized OFW | adversarial online multiclass logistic regression · MNIST and CIFAR10 | per-round computational cost | Hurt |
| Use ORGFW instead of the compared methods | constrained one-hidden-layer neural network optimization · MNIST and CIFAR10 | suboptimality | Helped |
| Use ORGFW instead of the compared methods | constrained one-hidden-layer neural network optimization · MNIST and CIFAR10 | training accuracy | Helped |
| Use ORGFW instead of the compared methods | constrained one-hidden-layer neural network optimization · MNIST and CIFAR10 | test accuracy | Helped |
| Use recursive stochastic gradient estimation | ORGFW · stochastic OCO | regret | Helped |
| Use recursive stochastic gradient estimation in the meta Frank-Wolfe framework | MORGFW · adversarial OCO | per-round stochastic gradient evaluations: \mathcal{O}(T^{3/2}) -> \mathcal{O}(T) | Helped |
Closest papers by meaning
- Projection-Free Online Convex Optimization with Time-Varying Constraints (ICML 2024 · similarity 0.90)
- Universal Online Convex Optimization with $1$ Projection per Round (NeurIPS 2024 · similarity 0.90)
- Revisiting Projection-Free Online Learning with Time-Varying Constraints (AAAI 2025 · similarity 0.89)
- Optimal Algorithms for Online Convex Optimization with Adversarial Constraints (NeurIPS 2024 · similarity 0.88)
- Improved Differentially Private and Lazy Online Convex Optimization: Lower Regret without Smoothness Requirements (ICML 2024 · similarity 0.88)
- Unified Projection-Free Algorithms for Adversarial DR-Submodular Optimization (ICLR 2024 · 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.