Research map / Continual learning / Online and replay-based
Adversarial Continual Learning
Sayna Ebrahimi, Franziska Meier, Roberto Calandra, Trevor Darrell, Marcus Rohrbach
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 |
|---|---|---|---|
| add adversarial discriminator D to shared and private modules (ablation) | ACL (S+P) · 20-Split miniImageNet | ACC: 50.15 -> 55.72 | Improved |
| add adversarial discriminator D to shared and private modules (ablation) | ACL (S+P) · 20-Split miniImageNet | BWT: -14.32 -> -0.12 | Improved |
| add orthogonality difference loss to shared and private modules (ablation) | ACL (S+P) · 20-Split miniImageNet | ACC: 50.15 -> 49.13 | No clear effect |
| add orthogonality difference loss to shared and private modules (ablation) | ACL (S+P) · 20-Split miniImageNet | BWT: -14.32 -> -3.99 | Improved |
| add difference loss to S+P+D (ablation) | ACL · 20-Split miniImageNet | ACC: 55.72 -> 57.66 | Improved |
| add one-sample-per-class replay buffer to S+P+D+Ldiff (ablation) | ACL · 20-Split miniImageNet | ACC: 57.66 -> 62.07 | Improved |
| add one-sample-per-class replay buffer to S+P+D+Ldiff (ablation) | ACL · 20-Split miniImageNet | BWT: -3.71 -> 0.00 | Improved |
| add discriminator rather than difference loss given replay buffer (ablation) | ACL + RB · 20-Split miniImageNet | ACC: 52.07 -> 60.28 | Improved |
| add shared module to private modules, discriminator and more components (shared only vs private only) (ablation) | ACL · 20-Split miniImageNet | ACC: 29.09 -> 21.19 | Hurt |
| add replay buffer to S+P (ablation) | ACL (S+P) · 20-Split miniImageNet | ACC: 50.15 -> 51.19 | Improved |
| increase replay samples per class from 1 to 13 | ER-RES · 20-Split miniImageNet | ACC: 40.21 -> 57.32 | Improved |
| increase replay samples per class from 1 to 13 | ACL · 20-Split miniImageNet | ACC: 62.07 -> 61.33 | No clear effect |
Closest papers by meaning
- A continual learning survey: Defying forgetting in classification tasks (TPAMI 2021 · similarity 0.91)
- Resurrecting Old Classes with New Data for Exemplar-Free Continual Learning (CVPR 2024 · similarity 0.90)
- Recall-Oriented Continual Learning with Generative Adversarial Meta-Model (AAAI 2024 · similarity 0.89)
- Continual Learning with Global Alignment (NeurIPS 2024 · similarity 0.89)
- Towards Understanding Catastrophic Forgetting in Two-layer Convolutional Neural Networks (ICML 2025 · similarity 0.89)
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.88)
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.
