Research map / Continual learning / Online and replay-based
Continual Reinforcement Learning with Complex Synapses
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 |
|---|---|---|---|
| Add Benna-Fusi synaptic dynamics to tabular Q-values | naive Q(λ) · 10x10 grid-world with alternating goal locations | time to relearn | Helped |
| Scale consolidation flow by eligibility traces (ablation) | Benna-Fusi Q(λ) · 10x10 grid-world with alternating goal locations | time to relearn | Helped |
| Add Benna-Fusi synaptic dynamics to network parameters | DQN · Cart-Pole and Catcher, alternating tasks | time to relearn | Helped |
| Add Benna-Fusi synaptic dynamics to network parameters | DQN · Cart-Pole, online single-task learning without replay | reward | Helped |
| Add Benna-Fusi synaptic dynamics to network parameters | DQN · Catcher, online single-task learning without replay | time to learn | Hurt |
| Use Benna-Fusi parameters across varied epoch lengths (ablation) | DQN · Cart-Pole and Catcher, alternating tasks | time to relearn | Helped |
| Use Benna-Fusi parameters while rotating across three tasks (ablation) | DQN · Cart-PoleLong, Catcher, and Cart-Pole | time to relearn | Helped |
| Use Benna-Fusi parameters instead of relying on a replay database (ablation) | DQN · Cart-Pole, online learning and small replay databases | reward | Helped |
| Prevent ε from decaying below a positive minimum (ablation) | Control DQN · Cart-Pole, online single-task learning | stable reward | No clear effect |
| Use Benna-Fusi parameters in larger DQN experiments | DQN · Arcade Learning Environment | performance | Hurt |
Closest papers by meaning
- Continual Learning Through Synaptic Intelligence (ICML 2017 · similarity 0.93)
- Temporal-Difference Variational Continual Learning (NeurIPS 2025 · similarity 0.87)
- Optimal Protocols for Continual Learning via Statistical Physics and Control Theory (ICLR 2025 · similarity 0.86)
- Predicting the Susceptibility of Examples to Catastrophic Forgetting (ICML 2025 · similarity 0.86)
- Mitigating Catastrophic Forgetting in Online Continual Learning by Modeling Previous Task Interrelations via Pareto Optimization (ICML 2024 · similarity 0.86)
- Unlocking the Power of Rehearsal in Continual Learning: A Theoretical Perspective (ICML 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.