Research map / Continual learning / Online and replay-based
BTSP-CAM: A Brain-Inspired Geometric Memory for Class-Incremental Learning
Zheng Zhang, Jiaye Yang, Qingjie Guo, Jiangrong Shen, Long Chen, Qi Xu
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 |
|---|---|---|---|
| add trace-gated plateau probability (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | average accuracy: 85.07 -> 86.82 | Improved |
| add trace-gated plateau probability (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | forgetting: 17.05 -> 14.33 | Improved |
| add collision-aware gating term (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | average accuracy: 85.11 -> 86.82 | Improved |
| add collision-aware gating term (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | old-to-new error: 14.8 -> 8.5 | Improved |
| add LTD (bidirectional flips, not only LTP) (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | average accuracy: 83.67 -> 86.82 | Improved |
| add LTD (bidirectional flips, not only LTP) (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | old-to-new error: 20.3 -> 8.5 | Improved |
| add CA1 competition layer (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | average accuracy: 82.34 -> 86.82 | Improved |
| add CA1 competition layer (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | decision margin: 0.68 -> 1.25 | Improved |
| add ridge consolidation (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | average accuracy: 75.81 -> 86.82 | Improved |
| add ridge consolidation (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (ablation) | decision margin: 0.41 -> 1.25 | Improved |
| add CA1 competition under 60% bit drop (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (partial cue) | old-class accuracy: 41.73 -> 68.11 | Improved |
| add LTD under 60% bit drop (ablation) | BTSP-CAM · CIFAR-100 B0 Inc5 (partial cue) | old-class accuracy: 50.91 -> 68.11 | Improved |
Closest papers by meaning
- Beyond Myopic Alignment: Lookahead Optimization for Online Class-Incremental Learning (CVPR 2026 · similarity 0.86)
- Hebbian and Gradient-based Plasticity Enables Robust Memory and Rapid Learning in RNNs (ICLR 2023 · similarity 0.86)
- Parameter-Masked Decoupled Optimization for Cross-Domain Class-Incremental Learning (ICML 2026 · similarity 0.86)
- Learning the Plasticity: Plasticity-Driven Learning Framework in Spiking Neural Networks (NeurIPS 2025 · similarity 0.84)
- The Geometric Inductive Bias of Grokking: Bypassing Phase Transitions via Architectural Topology (NeurIPS 2026 · similarity 0.84)
- Long-Tail Class Incremental Learning via Independent Sub-prototype Construction (CVPR 2024 · similarity 0.84)
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.
