Research map / Probabilistic and scientific modeling / Causal discovery
Sequential Causal Imitation Learning with Unobserved Confounders
Daniel Kumor, Junzhe Zhang, Elias Bareinboim
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 |
|---|---|---|---|
| Use all observed covariates instead of sequential π-Backdoor contexts | Behavioral cloning · Simulation 1, graph 1 | absolute imitation error: 0.04\pm0.04\% vs 0.13\pm0.18\% | Hurt |
| Use observed-parent contexts instead of sequential π-Backdoor contexts | Behavioral cloning · Simulation 1, graph 2 | absolute imitation error: 0.05\pm 0.03\% vs 0.20\pm0.25\% | Hurt |
| Use single-action π-Backdoor contexts instead of sequential π-Backdoor contexts | Behavioral cloning · Simulation 1, graph 3 | imitability | Hurt |
| Use sequential π-Backdoor contexts instead of non-causal context selection | Behavioral cloning · Simulation 1, graph 3 | absolute imitation error: 0.04\pm 0.03\% vs 0.27\pm0.40\%; 0.04\pm 0.03\% vs 0.26\pm0.39\% | Improved |
| Use non-causal contexts rather than checking imitability with the causal graph | Behavioral cloning · Simulation 1, graph 4 | absolute imitation error: 0.19\pm0.29\% | Hurt |
Closest papers by meaning
- Causal Imitability Under Context-Specific Independence Relations (NeurIPS 2023 · similarity 0.91)
- Causal Imitation for Markov Decision Processes: a Partial Identification Approach (NeurIPS 2024 · similarity 0.89)
- Is Behavior Cloning All You Need? Understanding Horizon in Imitation Learning (NeurIPS 2024 · similarity 0.87)
- Imitation Learning from Observation with Automatic Discount Scheduling (ICLR 2024 · similarity 0.85)
- A Smooth Sea Never Made a Skilled SAILOR: Robust Imitation via Learning to Search (NeurIPS 2025 · similarity 0.85)
- Learning When to Stop: Selective Imitation Learning Under Arbitrary Dynamics Shift (NeurIPS 2026 · 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.
