Spectral Representation Learning for Conditional Moment Models
This addresses a key issue in causal inference and economics for researchers using flexible models, though it appears incremental as it builds on known spectral decomposition approaches.
The paper tackles the challenge of validating ill-posedness measures in nonparametric conditional moment models by proposing a method to automatically learn representations with controlled ill-posedness, showing L2 consistency and promising performance on high-dimensional, semi-synthetic data.
Many problems in causal inference and economics can be formulated in the framework of conditional moment models, which characterize the target function through a collection of conditional moment restrictions. For nonparametric conditional moment models, efficient estimation often relies on preimposed conditions on various measures of ill-posedness of the hypothesis space, which are hard to validate when flexible models are used. In this work, we address this issue by proposing a procedure that automatically learns representations with controlled measures of ill-posedness. Our method approximates a linear representation defined by the spectral decomposition of a conditional expectation operator, which can be used for kernelized estimators and is known to facilitate minimax optimal estimation in certain settings. We show this representation can be efficiently estimated from data, and establish L2 consistency for the resulting estimator. We evaluate the proposed method on proximal causal inference tasks, exhibiting promising performance on high-dimensional, semi-synthetic data.