Research map / Continual learning / Online and replay-based
Self-Evolved Dynamic Expansion Model for Task-Free Continual Learning
Fei Ye, Adrian G. Bors
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 DEKMM soft-mask knowledge transfer (ablation) | SEDEM · Split CIFAR10 | accuracy: 52.98 -> 55.27 | Improved |
| add DEKMM soft-mask knowledge transfer (ablation) | SEDEM · Split CIFAR100 | accuracy: 22.74 -> 24.85 | Improved |
| add DEKMM soft-mask knowledge transfer (ablation) | SEDEM · Split MNIST | accuracy: 97.42 -> 98.35 | Improved |
| use novelty-aware sample selection instead of no selection (ablation) | SEDEM · Split CIFAR10 | accuracy: 50.14 -> 55.27 | Improved |
| use novelty-aware sample selection instead of CoPE selection (ablation) | SEDEM · Split CIFAR10 | accuracy: 50.82 -> 55.27 | Improved |
| use novelty-aware sample selection instead of reservoir sampling (ablation) | SEDEM · Split CIFAR100 | accuracy: 22.97 -> 24.85 | Improved |
| increase expansion threshold beta (ablation) | SEDEM · Split MNIST | accuracy and number of experts | Improved |
| use diversity-based expansion with the same number of experts as baselines (ablation) | SEDEM vs CNDPM, Dynamic-OCM · Split CIFAR10 | accuracy | Improved |
| SEDEM versus CNDPM on mixed-domain streams with fewer parameters | dynamic expansion · Split M-S | accuracy: 47.64 -> 56.62 | Improved |
Closest papers by meaning
- Online Task-Free Continual Learning via Dynamic Expansionable Memory Distribution (CVPR 2025 · similarity 0.90)
- Dynamic Expansion Diffusion Learning for Lifelong Generative Modelling (AAAI 2025 · similarity 0.89)
- Self-Expansion of Pre-trained Models with Mixture of Adapters for Continual Learning (CVPR 2025 · similarity 0.89)
- Model Sensitivity Aware Continual Learning (NeurIPS 2024 · similarity 0.87)
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.87)
- Scalable Language Model with Generalized Continual Learning (ICLR 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.
