Computational Identifiability
Not on the map. Why?
Abstract (arXiv)
Identification conditions describe the computability of a target query or parameter of interest as a function of the type and amount of information available. In causal identification, this information is often expressed in the form of a causal graph, and data are observed or collected for some subset of variables in the graph. Target queries may be for a single effect alone or for a class of effects in a given model. The derivation of an identification algorithm then defines mathematically the process by which the desired causal effect(s) can be uniquely determined, theoretically, in expectation. Identifiability in expectation, or 'theoretical identifiability,' generally assumes asymptotic properties, infinite data, or other mathematically idealized conditions. In this paper, we explore a fundamental distinction between this theoretical, idealized notion of identifiability and a proposed alternative that is computation-bound. The framework we propose - 'computational identifiability' - is to instead define a finite computational search procedure for an empirical estimator. If this process finds an estimator empirically, within a desired error tolerance, then identifiability is satisfied, conditional on the specified assumptions of the search (i.e., a prior distribution over the parameters) and conditional on the search procedure itself. Through several experiments, we demonstrate how this framework allows us to answer fine-grained, practical identification questions, such as identification with small finite samples, with ambiguous graphical criteria, with mixed observational-interventional data, and across counterfactual data and estimands. Code is available at https://github.com/lbynum/metadentify.
Key facts. Cited by 1 paper in the map (data through 13 Oct 2026). Builds on 26 of its 52 references that are in the map. First posted on arXiv 8 Jun 2026.
Builds on
26 of 52 references.
References · 26
- 2018Conditional Neural Processescited by 423
- 2018Neural Processescited by 309
- 2016Identifying Causal Effects With Proxy Variables of an Unmeasured Confoundercited by 139
- 2026TabICLv2: A better, faster, scalable, and open tabular foundation modelcited by 124
- 2021The Causal-Neural Connection: Expressiveness, Learnability, and Inferencecited by 97
- 2020An Introduction to Proximal Causal Learningcited by 94
- 2013Counterfactual Probabilities: Computational Methods, Bounds and Applicationscited by 67
- 2016Deep Survival Analysiscited by 61
- 2022Neural Causal Models for Counterfactual Identification and Estimationcited by 55
- 2025Do-PFN: In-Context Learning for Causal Effect Estimationcited by 46
- 2025CausalPFN: Amortized Causal Effect Estimation via In-Context Learningcited by 44
- 2021Nested Counterfactual Identification from Arbitrary Surrogate Experimentscited by 38
14 more
- 2025Foundation Models for Causal Inference via Prior-Data Fitted Networkscited by 30
- 2013Probabilistic Evaluation of Sequential Plans from Causal Models with Hidden Variablescited by 23
- 2019A Potential Outcomes Calculus for Identifying Conditional Path-Specific Effectscited by 19
- 2020A General Method for Deriving Tight Symbolic Bounds on Causal Effectscited by 18
- 2025Black Box Causal Inference: Effect Estimation via Meta Predictioncited by 13
- 2025Estimating Interventional Distributions with Uncertain Causal Graphs through Meta-Learningcited by 11
- 2024Demystifying amortized causal discovery with transformerscited by 10
- 2021Causal Estimation with Functional Confounderscited by 9
- 2023Partial Identification of Causal Effects Using Proxy Variablescited by 7
- 2023Adaptive Conditional Quantile Neural Processescited by 3
- 2024Consistency of Neural Causal Partial Identificationcited by 3
- 2026Amortizing Causal Sensitivity Analysis via Prior Data-Fitted Networkscited by 3
- 2026IV-ICL: Bounding Causal Effects with Instrumental Variables via In-Context Learningcited by 3
- 2026Causal Identification from Counterfactual Data: Completeness and Bounding Resultscited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.