Research map / Continual learning / Online and replay-based
Lifelong Neural Predictive Coding: Learning Cumulatively Online without Forgetting
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 S-NCN instead of the strongest listed baseline, VCL | continual classification · Split MNIST | ACC: 0.980 ± 0.210 -> 0.981 ± 0.300 | Helped |
| Use S-NCN instead of the strongest listed baseline, VCL | continual classification · Split Fashion MNIST | ACC: 0.980 ± 0.20 -> 0.982 ± 0.400 | Helped |
| Use S-NCN instead of the strongest listed baseline, VCL | continual classification · Split Not-MNIST | ACC: 0.953 ± 0.003 -> 0.957 ± 0.400 | Helped |
| Use FNBG-driven lateral inhibition and task-dependent contexts | S-NCN · continual learning benchmarks and task streams | memory retention | Helped |
| Use separate learnable feedback weights rather than symmetric weights | S-NCN · online generative modeling | reconstruction ability | Helped |
| Use separate learnable feedback weights rather than symmetric weights | S-NCN · classification | accuracy | No clear effect |
| Use fixed random backward synapses rather than learnable feedback weights | S-NCN · continual classification | ACC: reduced ACC by only a few points | Hurt |
Closest papers by meaning
- Sparse Coding in a Dual Memory System for Lifelong Learning (AAAI 2023 · similarity 0.90)
- Predictive Attractor Models (NeurIPS 2024 · similarity 0.88)
- A Stable, Fast, and Fully Automatic Learning Algorithm for Predictive Coding Networks (ICLR 2024 · similarity 0.87)
- When to Learn and When to Stop: Quitting at the Optimal Time (Student Abstract) (AAAI 2025 · similarity 0.87)
- Do Your Best and Get Enough Rest for Continual Learning (CVPR 2025 · similarity 0.87)
- PhiNets: Brain-inspired Non-contrastive Learning Based on Temporal Prediction Hypothesis (ICLR 2025 · 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.