Amortizing Causal Sensitivity Analysis via Prior Data-Fitted Networks

TL;DR

This paper introduces an amortized in-context learning method for causal sensitivity analysis that significantly speeds up computation by leveraging prior data and neural networks, unlike traditional per-instance methods.

Contribution

It proposes the first foundation model approach for causal sensitivity analysis, enabling rapid bounds estimation across datasets and sensitivity levels.

Findings

Achieves orders of magnitude faster test-time computation than traditional methods.

Develops a general prior-data construction applicable to various sensitivity models.

Recovers the full Pareto frontier of solutions under standard conditions.

Abstract

Causal sensitivity analysis aims to provide bounds for causal effect estimates in the presence of unobserved confounding. However, existing methods for causal sensitivity analysis are per-instance procedures, meaning that changes to the dataset, causal query, sensitivity level, or treatment require new computation. Here, we instead present an in-context learning approach. Specifically, we propose an amortized approach to causal sensitivity analysis based on prior-data fitted networks. A key challenge is that the sensitivity bounds are not directly available when sampling training data. To address this, we develop a general prior-data construction that is applicable across the class of generalized treatment sensitivity models. Our construction involves a Lagrangian scalarization of the objective to generate training labels for the bounds through a tradeoff between causal effect min/max-imization and sensitivity model violation, which avoids model-specific analytical derivations. We further show that, under standard convexity and linearity conditions, our objective recovers the full Pareto frontier of solutions. Empirically, we demonstrate our amortized approach across various datasets, causal queries, and sensitivity levels, where our approach achieves a test-time computation that is orders of magnitude faster than per-instance methods. To the best of our knowledge, ours is the first foundation model for in-context learning for causal sensitivity analysis.