Research map / Continual learning / Online and replay-based
Reinforced Continual Learning
Ju Xu, Zhanxing Zhu
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 |
|---|---|---|---|
| Use RCL instead of PGN | Continual learning model · Incremental CIFAR-100 | number of parameters: 42% | Improved |
| Use RCL instead of DEN | Continual learning model · Incremental CIFAR-100 | number of parameters: 53% | Improved |
| Use RCL instead of PGN and DEN | Continual learning model · MNIST Permutations, MNIST Mix, and Incremental CIFAR-100 | test accuracy | Improved |
| Dynamically expand the network rather than keep a fixed-size architecture | Continual learning model · MNIST Permutations, MNIST Mix, and Incremental CIFAR-100 | test accuracy | Improved |
| Use RCL rather than methods without network expansion | Continual learning model · Sequential task learning | accuracy on the first task | Improved |
| Use RCL rather than PGN | Continual learning model · MNIST Permutations | training time (seconds): 452 -> 34583 | Hurt |
| Use RCL rather than PGN | Continual learning model · MNIST Mix | training time (seconds): 451 -> 23626 | Hurt |
| Use RCL rather than PGN | Continual learning model · Incremental CIFAR-100 | training time (seconds): 167 -> 3936 | Hurt |
| Increase the reward coefficient alpha (ablation) | RCL · MNIST Permutations, MNIST Mix, and Incremental CIFAR-100 | number of model parameters | Improved |
| Increase the reward coefficient alpha (ablation) | RCL · MNIST Permutations, MNIST Mix, and Incremental CIFAR-100 | test accuracy | Hurt |
Closest papers by meaning
- Adaptive Progressive Continual Learning (TPAMI 2021 · similarity 0.93)
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.89)
- Improving Continual Learning Performance and Efficiency with Auxiliary Classifiers (ICML 2025 · similarity 0.89)
- Provable Contrastive Continual Learning (ICML 2024 · similarity 0.89)
- Self-Composing Policies for Scalable Continual Reinforcement Learning (ICML 2024 · similarity 0.88)
- Learning to Continually Learn with the Bayesian Principle (ICML 2024 · 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.
