Research map / Reinforcement learning / Visual and representation RL
An Inference-Based Policy Gradient Method for Learning Options
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 |
|---|---|---|---|
| infer options as latent variables (IOPG) instead of option-critic training | IOPG vs OC · MuJoCo continuous control | learning speed | Helped |
| use IOPG instead of A3C | IOPG vs A3C · Hopper, HalfCheetah, Swimmer | learning speed | No clear effect |
| use IOPG instead of A3C | IOPG vs A3C · Walker2d-v1 | learning speed | Helped |
| add Generalized Advantage Estimation to IOAC (ablation) | IOAC · MuJoCo environments | return | No clear effect |
| increase available options from 2 to 4 or 8 (ablation) | IOPG · Walker2d-v1 | return: 2 -> 4 options | Helped |
| increase available options from 4 to 8 (ablation) | IOPG · Walker2d-v1 | return: 4 -> 8 options | No clear effect |
| infer options instead of augmenting state space with options | IOPG · MuJoCo | learning speed | Helped |
Closest papers by meaning
- The Option-Critic Architecture (AAAI 2017 · similarity 0.89)
- Learning With Options That Terminate Off-Policy (AAAI 2018 · similarity 0.89)
- Flexible Option Learning (NeurIPS 2021 · similarity 0.88)
- Unveiling Options with Neural Network Decomposition (ICLR 2024 · similarity 0.87)
- Discovery of Options via Meta-Learned Subgoals (NeurIPS 2021 · similarity 0.87)
- Decision-Focused Learning with Directional Gradients (NeurIPS 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.