Research map / Continual learning / Online and replay-based
Continual Learning with Adaptive Weights (CLAW)
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 CLAW's data-driven adaptive neuron weights instead of LTG | continual learning models · Permuted MNIST (task 10) | classification accuracy: 99.2 ± 0.1 % vs 98.7 ± 0.3 % | Helped |
| Use CLAW's data-driven adaptive neuron weights instead of LTG | continual learning models · Split MNIST (task 5) | classification accuracy: 99.1 ± 0.2 % vs 98.3 ± 0.3 % | Helped |
| Use CLAW's data-driven adaptive neuron weights instead of LTG | continual learning models · Split notMNIST (task 5) | classification accuracy: 98.4 ± 0.2 % vs 97.4 ± 0.3 % | Helped |
| Use CLAW's data-driven adaptive neuron weights instead of LTG | continual learning models · Split Fashion-MNIST (task 5) | classification accuracy: 92.5 ± 0.2 % vs 91.1 ± 0.3 % | Helped |
| Use CLAW's data-driven adaptive neuron weights instead of LTG | continual learning models · Omniglot (task 50) | classification accuracy: 84.6 ± 0.3 % vs 83.5 ± 0.3 % | Helped |
| Use CLAW's data-driven adaptive neuron weights instead of LTG | continual learning models · CIFAR-100 (task 20) | classification accuracy: 95.6 ± 0.3 % vs 85.6 ± 0.5 % | Helped |
| Use one general maximum-adaptation value instead of task-specific values (ablation) | CLAW · Six continual-learning experiments | classification accuracy | Hurt |
| Always activate the binary adaptation decision (ablation) | CLAW · Six continual-learning experiments | classification accuracy | Hurt |
| Never adapt neuron contributions (ablation) | CLAW · Six continual-learning experiments | classification accuracy | Hurt |
Closest papers by meaning
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.89)
- TinySubNets: An Efficient and Low Capacity Continual Learning Strategy (AAAI 2025 · similarity 0.88)
- Learning to Continually Learn with the Bayesian Principle (ICML 2024 · similarity 0.88)
- PaLoRA: Paced Low-Rank Adaptation for Continual Learning (NeurIPS 2026 · similarity 0.88)
- Disentangling and mitigating the impact of task similarity for continual learning (NeurIPS 2024 · similarity 0.88)
- STAR: Stability-Inducing Weight Perturbation for Continual Learning (ICLR 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.