| { |
| "field_id": "F029", |
| "field": "causal inference", |
| "directions": [ |
| { |
| "direction_id": "G0", |
| "name": "Causal structure discovery", |
| "scope": "Recover causal directions, relationships, or graphs from observational, temporal, or interventional evidence, and address structural ambiguity caused by latent variables. Individual branches need not use all these forms of evidence together." |
| }, |
| { |
| "direction_id": "G1", |
| "name": "Identification of causal representations", |
| "scope": "Identify variables, factors, and representations from complex experiences or observations that can be given causal relationships and interventional meaning." |
| }, |
| { |
| "direction_id": "G2", |
| "name": "Causal identifiability and bounds", |
| "scope": "Determine what available distributions and assumptions establish about causal effects or counterfactual quantities: provide identification conditions and formulas, or bounds and sensitivity ranges when point identification is not possible." |
| }, |
| { |
| "direction_id": "G3", |
| "name": "Treatment-effect estimation and pathways", |
| "scope": "Use available data to estimate effects caused by interventions, including population, conditional, individual, and effects across time, and potentially decompose effects propagated through mediators, processes, or responding agents." |
| }, |
| { |
| "direction_id": "G4", |
| "name": "Individual counterfactual and response-type inference", |
| "scope": "Infer unobserved outcomes, outcome distributions, or potential-outcome patterns for the same individual under specified hypothetical changes, rather than reporting only population treatment effects." |
| }, |
| { |
| "direction_id": "G5", |
| "name": "Causal policy learning and evaluation", |
| "scope": "Use causal evidence to assess the consequences of action rules or select rules that improve outcomes, encompassing both optimization of rules for deployment and counterfactual evaluation of specified rules." |
| }, |
| { |
| "direction_id": "G6", |
| "name": "Causal experimental design and inference", |
| "scope": "Design assignments and analyze randomized or adaptive experiments to obtain valid, precise causal evidence." |
| }, |
| { |
| "direction_id": "G7", |
| "name": "Effect transport and evidence fusion", |
| "scope": "Combine evidence and infer effects for a target population or setting when trials, observational studies, sites, or data holders differ." |
| }, |
| { |
| "direction_id": "G8", |
| "name": "Causal fairness assessment and design", |
| "scope": "Assess or design predictions and action allocations with a focus on the causal effects of protected attributes, fairness criteria, or disparity constraints." |
| }, |
| { |
| "direction_id": "G9", |
| "name": "Causally invariant prediction", |
| "scope": "Find or construct predictive regularities that remain valid under relevant environmental changes and counterfactual interventions." |
| }, |
| { |
| "direction_id": "G10", |
| "name": "Attribution of realized causes", |
| "scope": "Explain which initiating disturbances, actions, or propagation pathways caused a particular realized outcome, fault, or anomaly." |
| }, |
| { |
| "direction_id": "G11", |
| "name": "Assessment of causal models and methods", |
| "scope": "Assess whether the models or methods underlying causal conclusions are credible, both by testing whether a given model is compatible with evidence and by developing evaluation resources and comparing discovery or estimation methods." |
| }, |
| { |
| "direction_id": "G12", |
| "name": "Causal-model foundations and semantics", |
| "scope": "Clarify the meanings of models, abstractions, interventions, and counterfactuals within causal frameworks, and the relationships between different formal frameworks." |
| } |
| ] |
| } |
|
|
