Research map / Continual learning / Online and replay-based
A Combinatorial Perspective on Transfer Learning
Jianan Wang, Eren Sezener, David Budden, Marcus Hutter, Joel Veness
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 NCTL instead of replay-free continual-learning baselines | Continual learning models · Split MNIST | accuracy: NCTL 95.07 ± 0.02; next-best replay-free baseline LwF 71.02 ± 1.26 | Improved |
| Use NCTL instead of replay-free continual-learning baselines | Continual learning models · Permuted MNIST | accuracy: NCTL 95.27 ± 0.00; next-best replay-free baseline MAS 94.08 ± 0.43 | Improved |
| Use NCTL instead of replay-based baselines | Continual learning models · Split MNIST and Permuted MNIST | accuracy: Split MNIST: NCTL 95.07 ± 0.02; RtF 97.31 ± 0.11. Permuted MNIST: NCTL 95.27 ± 0.00; RtF 97.06 ± 0.02 | Mixed |
| Train on more distractor tasks before returning to the initial task | NCTL · Free Split MNIST | instantaneous log-loss | No clear effect |
| Train on distractor tasks before returning to the initial task | NCTL · Free Split MNIST | instantaneous log-loss | Improved |
| Increase the number of source tasks | NCTL · Free Split MNIST held-out target tasks | instantaneous log-loss | Improved |
| Use NCTL rather than GLN or MLP | Online prediction models · Electricity (Elec2-3) | accuracy: 86.37% | Improved |
Closest papers by meaning
- Efficient Continual Learning with Modular Networks and Task-Driven Priors (ICLR 2021 · similarity 0.90)
- Learning curves for continual learning in neural networks: Self-knowledge transfer and forgetting (ICLR 2022 · similarity 0.89)
- Sequential Mastery of Multiple Visual Tasks: Networks Naturally Learn to Learn and Forget to Forget (CVPR 2020 · similarity 0.88)
- A Probabilistic Framework for Modular Continual Learning (ICLR 2024 · similarity 0.88)
- Disentangling and mitigating the impact of task similarity for continual learning (NeurIPS 2024 · similarity 0.87)
- Breaking Neural Network Scaling Laws with Modularity (ICLR 2025 · 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.
