Research map / Optimization / Nonconvex and smooth optimization
Data Summarization via Bilevel Optimization
Zalán Borsos, Mojmír Mutný, Marco Tagliasacchi, Andreas Krause
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| build coreset jointly for WideResNet-16-4 and VGG16 (ablation) | BiCo · CIFAR-10, VGG16 | test accuracy: 93.66 -> 93.88 | Improved |
| build coreset jointly for WideResNet-16-4 and VGG16 (ablation) | BiCo · CIFAR-10, MobileNetV2 (transfer) | test accuracy: 93.65 -> 93.79 | Improved |
| build coreset jointly for WideResNet-16-4 and VGG16 (ablation) | BiCo · SVHN, VGG16 | test accuracy: 95.43 -> 95.75 | Improved |
| build coreset jointly for WideResNet-16-4 and VGG16 (ablation) | BiCo · SVHN, MobileNetV2 (transfer) | test accuracy: 95.53 -> 95.76 | Improved |
| build coreset jointly for WideResNet-16-4 and VGG16 (ablation) | BiCo · CIFAR-10, WideResNet-16-4 | test accuracy: 95.24 -> 95.18 | Hurt |
| build coreset jointly for WideResNet-16-4 and VGG16 (ablation) | BiCo · SVHN, WideResNet-16-4 | test accuracy: 96.97 -> 96.88 | Hurt |
| use BiCo coreset instead of uniform sampling (transferred) | WideResNet-16-4 coreset size 23000 · CIFAR-10, VGG16 | test accuracy: 91.49 -> 93.66 | Improved |
| use BiCo coreset instead of uniform sampling (transferred) | WideResNet-16-4 coreset size 23000 · SVHN, MobileNetV2 | test accuracy: 94.22 -> 95.53 | Improved |
| use BiCo replay memory instead of uniform sampling | continual learning · SMNIST, memory 50 | accuracy: 85.23 -> 91.61 | Improved |
| use BiCo replay memory instead of uniform sampling | continual learning · SMNIST, memory 200 | accuracy: 95.08 -> 97.01 | Improved |
| use BiCo merge-reduce buffer instead of reservoir sampling | streaming · SMNIST, memory 50 | accuracy: 83.90 -> 85.32 | Improved |
| add sample weights (regularization-based weighted coreset) to unweighted selection | BiCo on binary logistic regression · LIBSVM binary datasets | coreset size to match full-data accuracy: 2% of the data vs twice as large | Improved |
Closest papers by meaning
- Bilevel Optimization under Unbounded Smoothness: A New Algorithm and Convergence Analysis (ICLR 2024 · similarity 0.85)
- TSP: A Two-Sided Smoothed Primal-Dual Method for Nonconvex Bilevel Optimization (ICML 2025 · similarity 0.85)
- A Nearly Optimal Single Loop Algorithm for Stochastic Bilevel Optimization under Unbounded Smoothness (ICML 2024 · similarity 0.84)
- Refined Coreset Selection: Towards Minimal Coreset Size under Model Performance Constraints (ICML 2024 · similarity 0.84)
- Universal Weak Coreset (AAAI 2024 · similarity 0.84)
- $\mathbb{D}^2$ Pruning: Message Passing for Balancing Diversity & Difficulty in Data Pruning (ICLR 2024 · 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.
