Research map / Continual learning / Online and replay-based
Sparse Coding in a Dual Memory System for Lifelong Learning
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 SCoMMER instead of CLS-ER | ResNet-18 · S-CIFAR10 Class-IL, buffer 200 | accuracy: 66.19 ±0.75 -> 69.19 ±0.61 | Helped |
| Use SCoMMER instead of CLS-ER | ResNet-18 · S-CIFAR100 Class-IL, buffer 200 | accuracy: 35.23 ±0.86 -> 40.25 ±0.05 | Helped |
| Use SCoMMER instead of CLS-ER | ResNet-18 · GCIL Unif, buffer 200 | accuracy: 25.06 ±0.81 -> 30.84 ±0.80 | Helped |
| Use SCoMMER instead of CLS-ER | ResNet-18 · S-CIFAR10 Class-IL, buffer 500 | accuracy: 75.22 ±0.71 -> 74.97 ±1.05 | Hurt |
| Use SCoMMER instead of CLS-ER | ResNet-18 · S-CIFAR10 Task-IL, buffer 500 | accuracy: 94.94 ±0.53 -> 94.36 ±0.06 | Hurt |
| Use SCoMMER instead of CLS-ER | ResNet-18 · S-CIFAR100 Class-IL, buffer 500 | accuracy: 47.63 ±0.61 -> 49.63 ±1.43 | Helped |
| Use SCoMMER instead of CLS-ER | ResNet-18 · S-CIFAR100 Task-IL, buffer 500 | accuracy: 73.78 ±0.86 -> 75.49 ±0.43 | Helped |
| Use SCoMMER instead of CLS-ER | ResNet-18 · GCIL Longtail, buffer 500 | accuracy: 28.63 ±0.68 -> 35.20 ±0.21 | Helped |
| Add semantic dropout (ablation) | SCoMMER · S-CIFAR10 Class-IL, buffer 200 | accuracy: 67.38 ±1.51 -> 69.19 ±0.61 | Helped |
| Add sparse activations to the variant without semantic dropout (ablation) | SCoMMER · S-CIFAR10 Class-IL, buffer 200 | accuracy: 61.88 ±2.43 -> 67.38 ±1.51 | Helped |
| Add long-term memory to the variant without semantic dropout (ablation) | SCoMMER · S-CIFAR10 Class-IL, buffer 200 | accuracy: 49.44 ±5.43 -> 67.38 ±1.51 | Helped |
| Use long-term memory rather than working memory | SCoMMER · S-CIFAR10 Class-IL, buffer 200 | accuracy: 58.03 ±5.17 -> 69.19 ±0.61 | Helped |
Closest papers by meaning
- Learning Bayesian Sparse Networks With Full Experience Replay for Continual Learning (CVPR 2022 · similarity 0.90)
- Lifelong Neural Predictive Coding: Learning Cumulatively Online without Forgetting (NeurIPS 2022 · similarity 0.90)
- Do Your Best and Get Enough Rest for Continual Learning (CVPR 2025 · similarity 0.87)
- InTAct: Interval-based Task Activation Consolidation for Continual Learning (NeurIPS 2026 · similarity 0.87)
- Task-aware Orthogonal Sparse Network for Exploring Shared Knowledge in Continual Learning (ICML 2024 · similarity 0.87)
- A Unified and General Framework for Continual Learning (ICLR 2024 · 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.