Research map / Optimization / Distributed and compressed
Communication-Efficient Distributed Stochastic AUC Maximization with Deep Neural Networks
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 more machines | CoDA · CIFAR10, CIFAR100, and ImageNet (positive ratio 71%) | iterations to target AUC | Helped |
| Use 16 machines | CoDA · CIFAR10, CIFAR100, and ImageNet (positive ratio 71%) | running time | Helped |
| Increase the communication interval to I=1024 (ablation) | CoDA with K=16 · ImageNet (positive ratio 71%) | running time | Helped |
| Increase the communication interval to I=64 (ablation) | CoDA with K=16 · CIFAR100 (positive ratio 71%) | running time | Helped |
| Increase the communication interval to I=64 (ablation) | CoDA with K=16 · CIFAR10 (positive ratio 71%) | running time | Helped |
| Increase the communication interval to I=512 (ablation) | CoDA with K=4 · CIFAR100 (positive ratio 71%) | iterations to target AUC | No clear effect |
| Increase the communication interval to I=512 (ablation) | CoDA with K=4 · CIFAR10 (positive ratio 71%) | iterations to target AUC | No clear effect |
| Increase the communication interval beyond the reported tolerance threshold (ablation) | CoDA with K=16 · ImageNet, CIFAR100, and CIFAR10 (positive ratio 71%) | testing AUC | Hurt |
| Use I>1 to skip communication (ablation) | CoDA · CIFAR10, CIFAR100, and ImageNet (positive ratio 71%) | iterations to target AUC | No clear effect |
| Use I>1 to skip communication (ablation) | CoDA · CIFAR10, CIFAR100, and ImageNet (positive ratio 71%) | running time | Helped |
| Use fewer machines (K=4 rather than K=16) (ablation) | CoDA · CIFAR10 and CIFAR100 (positive ratio 71%) | maximum communication interval without dramatic performance drop: I=512; I=64 | Helped |
Closest papers by meaning
- Stochastic AUC Maximization with Deep Neural Networks (ICLR 2020 · similarity 0.92)
- Gradient Sparsification for Communication-Efficient Distributed Optimization (NeurIPS 2018 · similarity 0.90)
- Distributed Optimization for Overparameterized Problems: Achieving Optimal Dimension Independent Communication Complexity (NeurIPS 2022 · similarity 0.89)
- Trading Redundancy for Communication: Speeding up Distributed SGD for Non-convex Optimization (ICML 2019 · similarity 0.89)
- Hybrid Decentralized Optimization: Leveraging Both First- and Zeroth-Order Optimizers for Faster Convergence (AAAI 2025 · similarity 0.87)
- Towards Faster Decentralized Stochastic Optimization with Communication Compression (ICLR 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.