Research map / Probabilistic and scientific modeling / Treatment effects
Learning Generalized Gumbel-max Causal Mechanisms
Guy Lorberbom, Daniel D. Johnson, Chris J. Maddison, Daniel Tarlow, Tamir Hazan
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 |
|---|---|---|---|
| Learn Gadget 2 for couplings over independent p and q | Categorical outcome coupling · Independent p and q, fixed linear reward | variance: 14.02 ± 0.01 -> 8.76 ± 0.03 | Improved |
| Learn Gadget 2 for couplings over mirrored p and q | Categorical outcome coupling · Mirrored p and q, fixed linear reward | variance: 18.84 ± 0.09 -> 13.47 ± 0.09 | Improved |
| Learn Gadget 1 for couplings over mirrored p and q | Categorical outcome coupling · Mirrored p and q, fixed linear reward | variance: 18.84 ± 0.09 -> 16.67 ± 0.45 | Improved |
| Learn Gadget 2 for couplings over fixed increasing/decreasing distributions | Categorical outcome coupling · Random monotonic reward | variance: 2.46 ± 0.35 -> 2.00 ± 0.30 | Mixed |
| Learn Gadget 1 for couplings over fixed increasing/decreasing distributions | Categorical outcome coupling · Random monotonic reward | variance: 2.46 ± 0.35 -> 1.42 ± 0.22 | Improved |
| Learn Gadget 2 for couplings over fixed increasing/decreasing distributions | Categorical outcome coupling · Non-monotonic reward | variance: 0.50 ± 0.06 -> 0.21 ± 0.03 | Improved |
| Learn Gadget 1 for couplings over fixed increasing/decreasing distributions | Categorical outcome coupling · Non-monotonic reward | variance: 0.50 ± 0.06 -> 0.26 ± 0.03 | Improved |
| Train the learned coupling on a narrow distribution of query pairs | Gadget 2 · Near-maximal coupling experiment | probability of matching outcomes | Improved |
| Learn Gadget 2 instead of using fixed sampling mechanisms | Sepsis management MDP · Counterfactual treatment effects | variance | Improved |
Closest papers by meaning
- Cyclic Counterfactuals under Shift–Scale Interventions (NeurIPS 2025 · similarity 0.89)
- Causal learning with the invariance principle (NeurIPS 2026 · similarity 0.89)
- Towards Characterizing Domain Counterfactuals for Invertible Latent Causal Models (ICLR 2024 · similarity 0.88)
- Counterfactual Identification Under Monotonicity Constraints (AAAI 2025 · similarity 0.88)
- Exogenous Isomorphism for Counterfactual Identifiability (ICML 2025 · similarity 0.87)
- Gumbel Counterfactual Generation From Language Models (ICLR 2025 · similarity 0.87)
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.
