Algorithmic Feature Highlighting for Human-AI Decision-Making

arXiv cs.LG / 4/27/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper studies algorithms that select a small, case-specific subset of features to highlight for human review, instead of directly outputting a single prediction or recommendation.
It models feature highlighting as a constrained information policy and shows that human interpretation differs sharply depending on whether the human agent accounts for the selection rule.
Optimizing highlighting for a sophisticated (rule-aware) agent is often computationally intractable, even in simple discrete/binary settings, while optimization for a naive agent can be tractable if maximum bandwidth is fixed.
A highlighting policy optimized for sophisticated agents can perform arbitrarily badly when used by naive agents, motivating robust and implementable designs that tolerate human misunderstandings.
The framework is illustrated with a calibrated empirical study using the American Housing Survey, arguing for context-specific highlighting to achieve practical human–algorithm complementarity.

Abstract

Human decision-makers often face choices about complex cases with many potentially relevant features, but limited bandwidth to inspect and integrate all available information. In such settings, we study algorithms that highlight a small subset of case-specific features for human consideration, rather than producing a single prediction or recommendation. We model highlighting as a constrained information policy that selects a small number of features to reveal. A central issue is how humans interpret the algorithm's choice of features: a sophisticated agent correctly conditions on the selection rule, while a naive agent updates only on revealed feature values and treats the selection event as exogenous. We show that optimizing highlighting for sophisticated agents can be computationally intractable, even in simple discrete and binary settings, whereas optimizing for naive agents is tractable as long as the maximal bandwidth is fixed. We also show that a highlighting policy that is optimal for sophisticated agents can perform arbitrarily poorly when deployed to naive agents, motivating robust, implementable alternatives. We illustrate our framework in a calibrated empirical exercise based on the American Housing Survey. Overall, our results establish the value of highlighting a context-specific set of features rather than a fixed one as a practically appealing and computationally feasible tool for achieving human-algorithm complementarity.