Research map / Continual learning / Online and replay-based
On the Effectiveness of Lipschitz-Driven Rehearsal in Continual Learning
Lorenzo Bonicelli, Matteo Boschini, Angelo Porrello, Concetto Spampinato, Simone Calderara
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 LiDER Lipschitz regularization on buffer samples | DER++ · Split CIFAR-100 (buffer 500) | Final avg accuracy: 37.13 -> 39.25 | Improved |
| add LiDER | DER++ · Split CUB-200 (buffer 400) | Final avg accuracy: 49.30 -> 57.90 | Improved |
| add LiDER | iCaRL · Split CIFAR-100 (buffer 500) | Final avg accuracy: 44.04 -> 47.02 | Improved |
| add LiDER | GDumb · Split CIFAR-100 (buffer 500) | Final avg accuracy: 09.28 -> 10.22 | Improved |
| add LiDER | ER-ACE · Split miniImageNet (buffer 2000) | Final avg accuracy: 22.60 -> 24.13 | Improved |
| apply LiDER to current-task examples instead of buffer examples (ablation) | DER++ · Split CIFAR-100 (buffer 500) | Final avg accuracy: 39.25 -> 34.78 | Hurt |
| use fixed Lipschitz targets instead of learned targets (ablation) | LiDER on DER++ · Split CIFAR-100 (buffer 500) | Final avg accuracy: 39.25 -> 36.42 | Hurt |
| add Stable SGD (ablation) | ER-ACE · Split CUB-200 (buffer 400) | Final avg accuracy: 41.83 -> 22.67 | Hurt |
| add online EwC (ablation) | ER-ACE · Split CUB-200 (buffer 400) | Final avg accuracy: 41.83 -> 48.34 | Improved |
| add LiDER under buffer label poisoning (p=0.25%) (ablation) | DER++ · Split CIFAR-100 (buffer 500) | Final avg accuracy: 28.74 -> 30.78 | Improved |
| add LiDER (ablation) | ER-ACE · Buffer guessing game | ROC-AUC of identifying buffer samples | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- GDumb — GDumb: A Simple Approach that Questions Our Progress in Continual Learning (ECCV 2020) · better on 2 of 2 shared results
- iCaRL — iCaRL: Incremental Classifier and Representation Learning (CVPR 2017) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Unlocking the Power of Rehearsal in Continual Learning: A Theoretical Perspective (ICML 2025 · similarity 0.90)
- Layerwise Proximal Replay: A Proximal Point Method for Online Continual Learning (ICML 2024 · similarity 0.88)
- Predicting the Susceptibility of Examples to Catastrophic Forgetting (ICML 2025 · similarity 0.88)
- STAR: Stability-Inducing Weight Perturbation for Continual Learning (ICLR 2025 · similarity 0.88)
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.88)
- Mitigating Catastrophic Forgetting in Online Continual Learning by Modeling Previous Task Interrelations via Pareto Optimization (ICML 2024 · similarity 0.87)
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.
