Research map / Continual learning / Online and replay-based
Exploring Example Influence in Continual Learning
Qing Sun, Fan Lyu, Fanhua Shang, Wei Feng, Liang Wan
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 influence-based rehearsal selection (RehSel) over random selection (ablation) | MetaSP · Split CIFAR-10 class-incremental, buffer 500 | final accuracy A_inf: 49.16 -> 50.10 | Improved |
| add RehSel (ablation) | MetaSP · Split CIFAR-100 class-incremental, buffer 500 | final accuracy A_inf: 18.96 -> 19.28 | Improved |
| add RehSel (ablation) | MetaSP · Split CIFAR-10 class-incremental, buffer 500 | first accuracy A_1 / A_m: 96.82 -> 96.81 (A_1) | Mixed |
| remove K-means clustering from rehearsal selection (ablation) | RehSel · Split CIFAR-10 class-incremental | final accuracy A_inf: 50.10 -> 38.50 | Hurt |
| remove influence ranking, cluster only (ablation) | RehSel · Split CIFAR-10 class-incremental | final accuracy A_inf: 50.10 -> 47.55 | Hurt |
| use only stability-aware or only plasticity-aware influence instead of fused (ablation) | MetaSP · Split CIFAR-10 | first/final accuracy | Mixed |
| use more pseudo-update iterations (ablation) | MetaSP · Split CIFAR-10 | accuracy | Improved |
| larger SP validation set (ablation) | MetaSP · Split CIFAR-10/100 | accuracy | Improved |
| apply MetaSP vs ER baseline | ER · Split CIFAR-10 class-incremental, buffer 500 | final accuracy A_inf: 40.45 -> 49.16 | Improved |
| apply MetaSP vs best baseline | GMED · Split Mini-Imagenet task-incremental, buffer 500 | final accuracy A_inf: 30.47 -> 32.59 | Improved |
| MetaSP influence approximation vs Neumann series | influence estimation · 1000 examples | true positives: 642 -> 650 | Improved |
Closest papers by meaning
- Predicting the Susceptibility of Examples to Catastrophic Forgetting (ICML 2025 · similarity 0.87)
- Pareto Continual Learning: Preference-Conditioned Learning and Adaption for Dynamic Stability-Plasticity Trade-off (AAAI 2025 · similarity 0.87)
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.87)
- STAR: Stability-Inducing Weight Perturbation for Continual Learning (ICLR 2025 · similarity 0.86)
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.86)
- Disentangling and mitigating the impact of task similarity for continual learning (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.
