Back to feed
arXiv cs.LG
arXiv cs.LG
6/19/2026
Computational Identifiability

Computational Identifiability

Short summary

A new framework called 'computational identifiability' redefines how to determine whether causal effects can be estimated from data. Unlike classical identifiability (which assumes infinite data), this approach uses finite, practical computational search—enabling causal inference with small samples and mixed observational-interventional data. Code is released on GitHub.

  • •New framework redefines causal identification as a computational search problem rather than purely theoretical property
  • •Enables causal inference with small finite samples, ambiguous graphical models, and mixed observational-interventional data
  • •Code available; validates on counterfactual estimands and practical real-world scenarios

Generated with AI, which can make mistakes.

#research-breakthrough
Read full article at arXiv cs.LG

Is this a good recommendation for you?

Explore more

Information Lattice Learning connects to
arXiv cs.LG
Beyond Shapley: Efficient Computation of Asymmetric Shapley Values
arXiv CS.AI
TTCD:Transformer Integrated Temporal Causal Discovery from Non-Stationary Time Series Data
arXiv cs.LG
CIExplainer++: Generating Causal and Interpretable Explanations for Graph Neural Networks
arXiv cs.LG