Research map / Reinforcement learning / Visual and representation RL
Optimizing Memory Placement using Evolutionary Graph Reinforcement Learning
Shauharda Khadka, Estelle Aflalo, Mattias Marder, Avrech Ben-David, Santiago Miret, Shie Mannor, Tamir Hazan, Hanlin Tang, Somdeb Majumdar
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 |
|---|---|---|---|
| Combine evolutionary search and policy-gradient learning with shared replay and policy migration | EGRL memory mapping · ResNet-50 | speedup: 1.28 vs 1.06 (EA), 0.72 (Greedy-DP), and 0.29 (PG) | Improved |
| Combine evolutionary search and policy-gradient learning with shared replay and policy migration | EGRL memory mapping · ResNet-101 | speedup: 1.78 vs 1.47 (EA), 1.27 (Greedy-DP), and 0.23 (PG) | Improved |
| Combine evolutionary search and policy-gradient learning with shared replay and policy migration | EGRL memory mapping · BERT | speedup: 1.66 vs 1.64 (EA), 0.67 (Greedy-DP), and 0.21 (PG) | Improved |
| Remove evolutionary search and use policy-gradient learning alone (ablation) | EGRL · ResNet-50 | speedup: 1.28 -> 0.29 | Hurt |
| Remove evolutionary search and use policy-gradient learning alone (ablation) | EGRL · ResNet-101 | speedup: 1.78 -> 0.23 | Hurt |
| Remove evolutionary search and use policy-gradient learning alone (ablation) | EGRL · BERT | speedup: 1.66 -> 0.21 | Hurt |
| Remove policy-gradient learning and use evolutionary search alone (ablation) | EGRL · ResNet-50 | speedup: 1.28 -> 1.06 | Hurt |
| Remove policy-gradient learning and use evolutionary search alone (ablation) | EGRL · ResNet-101 | speedup: 1.78 -> 1.47 | Hurt |
| Remove policy-gradient learning and use evolutionary search alone (ablation) | EGRL · BERT | speedup: 1.66 vs 1.64 | No clear effect |
| Use layer-wise greedy decisions with a conditional-independence assumption | Greedy-DP memory mapping · BERT | speedup: 0.67 | Hurt |
Closest papers by meaning
- Reinforced Genetic Algorithm Learning for Optimizing Computation Graphs (ICLR 2020 · similarity 0.90)
- Evolution-Guided Policy Gradient in Reinforcement Learning (NeurIPS 2018 · similarity 0.87)
- ESSAM: A Novel Competitive Evolution Strategies Approach to Reinforcement Learning for Memory Efficient LLMs Fine-Tuning (NeurIPS 2026 · similarity 0.87)
- Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks (NeurIPS 2026 · similarity 0.86)
- Mem-T: Densifying Rewards for Long-Horizon Memory Agents (ICML 2026 · similarity 0.86)
- Reinforcement Learning within Tree Search for Fast Macro Placement (ICML 2024 · similarity 0.85)
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.
