Research map / Learning theory and algorithms / Prediction and decision losses
Joint Training of Deep Ensembles Fails Due to Learner Collusion
Alan Jeffares, Tennison Liu, Jonathan Crabbé, Mihaela van der Schaar
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 |
|---|---|---|---|
| train ensemble jointly on ensemble loss instead of independently | ResNet-18 ensemble · ImageNet val | top-1 accuracy: 68.10 -> 62.74 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | ResNet-34 ensemble · ImageNet val | top-1 accuracy: 73.11 -> 68.91 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | DenseNet ensemble · ImageNet val | top-1 accuracy: 70.62 -> 60.81 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | ShuffleNet V2 0.5x ensemble · ImageNet val | top-1 accuracy: 61.00 -> 53.41 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | ShuffleNet V2 1x ensemble · ImageNet val | top-1 accuracy: 68.56 -> 61.96 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | SqueezeNet V1.1 ensemble · ImageNet val | top-1 accuracy: 56.25 -> 42.69 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | MobileNet V3-small ensemble · ImageNet val | top-1 accuracy: 67.43 -> 57.34 | Hurt |
| train ensemble jointly on ensemble loss instead of independently | MobileViT ensemble · ImageNet val | top-1 accuracy: 61.29 -> 54.69 | Hurt |
| increase beta toward 1 (more joint training) (ablation) | ResNet ensembles · CIFAR-10/100 | test accuracy | Hurt |
| increase beta toward 1 (more joint training) (ablation) | ResNet-18 base learners · ImageNet | validation accuracy | Hurt |
Closest papers by meaning
- Theoretical Limitations of Ensembles in the Age of Overparameterization (ICML 2025 · similarity 0.87)
- Ensemble Pruning for Out-of-distribution Generalization (ICML 2024 · similarity 0.87)
- No Free Lunch from Random Feature Ensembles: Scaling Laws and Near-Optimality Conditions (ICML 2025 · similarity 0.86)
- On the Diversity of Adversarial Ensemble Learning (ICML 2025 · similarity 0.86)
- The Disparate Benefits of Deep Ensembles (ICML 2025 · similarity 0.86)
- Sharpness-diversity tradeoff: improving flat ensembles with SharpBalance (NeurIPS 2024 · similarity 0.86)
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.
