Research map / Continual learning / Online and replay-based
Formalizing the Generalization-Forgetting Trade-off in Continual Learning
Krishnan Raghavan, Prasanna Balaprakash
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 |
|---|---|---|---|
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · split-MNIST ITL | retained accuracy: 97.73 \pm 0.03 -> 99.52\pm0.07 | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · split-MNIST IDL | retained accuracy: 96.43 \pm0.29 -> 98.71 \pm\textbf{0.06} | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · split-MNIST ICL | retained accuracy: 91.88\pm0.55 -> 97.32\pm 0.17 | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · permuted-MNIST ITL | retained accuracy: 96.16 \pm 0.03 -> 97.41 \pm 0.01 | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · permuted-MNIST IDL | retained accuracy: 96.08 \pm 0.06 -> 97.51 \pm 0.05 | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · permuted-MNIST ICL | retained accuracy: 95.96\pm0.06 -> 97.61\pm\textbf{0.01} | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · split-CiFAR100 ITL | retained accuracy: 69.17 \pm0.12 -> 81.82 \pm \textbf{0.17} | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · split-CiFAR100 IDL | retained accuracy: 51.82\pm0.19 -> 62.11 \pm \textbf{0.00} | Improved |
| fix the task-discrepancy player's play, removing the game (ablation) | BCL · split-CiFAR100 ICL | retained accuracy: 52.82 \pm 0.01 -> 69.27 \pm \textbf{0.03} | Improved |
Closest papers by meaning
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.90)
- STAR: Stability-Inducing Weight Perturbation for Continual Learning (ICLR 2025 · similarity 0.90)
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.89)
- Pareto Continual Learning: Preference-Conditioned Learning and Adaption for Dynamic Stability-Plasticity Trade-off (AAAI 2025 · similarity 0.89)
- Advancing Prompt-Based Methods for Replay-Independent General Continual Learning (ICLR 2025 · similarity 0.89)
- Mitigating Catastrophic Forgetting in Online Continual Learning by Modeling Previous Task Interrelations via Pareto Optimization (ICML 2024 · similarity 0.89)
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.
