Research map / Reinforcement learning / Policy gradient and actor-critic
Learning With Options That Terminate Off-Policy
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 longer behavior options by decreasing ζ (ablation) | Q(β) · 19-state random walk | sum error | Helped |
| use longer behavior options by decreasing ζ (ablation) | plain intra-option · 19-state random walk | sum error | Hurt |
| learn target terminations off-policy with Q(β), using long behavior options | Q(β) · 19-state random walk | value estimation accuracy | Helped |
| increase target termination β (ablation) | Q(β) · Modified Cliffwalk | performance | Helped |
| decouple target termination β from behavior termination ζ | Q(β) · Modified Cliffwalk | performance | Helped |
| increase target termination β (ablation) | Q(β) · Pinball | performance | Mixed |
| use longer behavior options by decreasing ζ (ablation) | Q(β) · Pinball | performance | Helped |
| decouple target termination β from behavior termination ζ | Q(β) · Pinball | performance | Helped |
| replace Q(β) with the off-policy-plain variant, learning with ζ=0 and evaluating with β (ablation) | Q(β) · Pinball | performance | No clear effect |
| use larger target termination β rather than smaller termination | greedy control with a fixed option set · theoretical control result | value | Helped |
Closest papers by meaning
- The Option-Critic Architecture (AAAI 2017 · similarity 0.89)
- An Inference-Based Policy Gradient Method for Learning Options (ICML 2018 · similarity 0.89)
- When Waiting Is Not an Option: Learning Options With a Deliberation Cost (AAAI 2018 · similarity 0.88)
- Skill or Luck? Return Decomposition via Advantage Functions (ICLR 2024 · similarity 0.87)
- Exploiting Action Impact Regularity and Exogenous State Variables for Offline Reinforcement Learning (Abstract Reprint) (AAAI 2024 · similarity 0.86)
- Seizing Serendipity: Exploiting the Value of Past Success in Off-Policy Actor-Critic (ICML 2024 · 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.