Research map / Reinforcement learning / Policy gradient and actor-critic
Evolved Policy Gradients
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 the EPG learned loss instead of PPO | test-time learning on randomized MuJoCo environments · RandomHopper, RandomWalker, RandomReacher, GoalAnt, Fetch | episodic return | Helped |
| use the EPG learned loss instead of PPO | DirectionalHopper and DirectionalHalfCheetah · direction-randomized tasks | episodic return | Helped |
| use RL2 instead of EPG | DirectionalHalfCheetah · DirectionalHalfCheetah | episodic return | Helped |
| use EPG instead of RL2 or MAML | test-time learning on goals outside the metatraining distribution · GoalAnt, negative x-axis goals | episodic return | Helped |
| use EPG instead of RL2 | test-time learning on goals outside the metatraining distribution · GoalAnt, negative x-axis goals | episodic return | Helped |
| use EPG instead of MAML | test-time learning on goals outside the metatraining distribution · GoalAnt, negative x-axis goals | episodic return | Helped |
| use EPG instead of MAML with 10 gradient updates | test-time learning on goals outside the metatraining distribution · GoalAnt, negative x-axis goals | learning speed | Helped |
| use EPG instead of RL2 | test-time learning on goals from the metatraining distribution · GoalAnt, positive x-axis goals | episodic return | No clear effect |
Closest papers by meaning
- Evolution-Guided Policy Gradient in Reinforcement Learning (NeurIPS 2018 · similarity 0.90)
- Improving Exploration in Evolution Strategies for Deep Reinforcement Learning via a Population of Novelty-Seeking Agents (NeurIPS 2018 · similarity 0.89)
- Evolving Reinforcement Learning Algorithms (ICLR 2021 · similarity 0.88)
- Discovering Temporally-Aware Reinforcement Learning Algorithms (ICLR 2024 · similarity 0.88)
- Can Learned Optimization Make Reinforcement Learning Less Difficult? (NeurIPS 2024 · similarity 0.87)
- EvoTest: Evolutionary Test-Time Learning for Self-Improving Agentic Systems (ICLR 2026 · 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.