Search papers, labs, and topics across Lattice.
This paper introduces Neural EXposure Interaction Search (NEXIS), a novel method for identifying heterogeneous treatment effects (HTE) by leveraging extensive pre-treatment measurements and scalable representations. By framing HTE identification as a Markov-blanket discovery problem, NEXIS achieves a causal characterization of treatment effects that is both interpretable and actionable. The application of NEXIS to anti-poverty programs in Africa demonstrates its ability to uncover previously unmeasured environmental modifiers, providing new insights for policy optimization.
Uncovering hidden environmental factors can transform anti-poverty strategies, revealing actionable insights that were previously overlooked.
Heterogeneous Treatment Effect (HTE) identification is crucial to explain the impact of an intervention and optimize our policies accordingly. Existing approaches trade expressivity for interpretability, but, if some active heterogeneity drivers are unmeasured, methods at both ends of this spectrum allow for spurious HTE characterization with no causal reading. In this work, we focus on controlled experiments and argue that an oracle HTE causal characterization via the latent interactors is now within reach, thanks to (i) more extensive pre-treatment measurements, i.e., multi-modal and multi-view, and (ii) scalable representations with minimal human supervision. We then re-frame HTE identification as a Markov-blanket discovery problem on a sufficient and aligned pre-treatment representation, and introduce Neural EXposure Interaction Search (NEXIS), an iterative procedure with provable and empirically validated consistent selection. We deploy NEXIS on two anti-poverty programs in Africa, augmenting each with satellite imagery capturing previously unmeasured environmental effect modifiers, leading to novel, interpretable and prescriptive guidelines to optimize the programs' next iterations.
