Relaxed Efficient Acquisition of Context and Temporal Features

2026-03-11

AI Summary

This work addresses a key challenge in biomedical prediction: jointly optimizing the selection of one-time contextual features and the adaptive acquisition of longitudinal features under limited cost constraints. The authors propose REACT, a novel framework that unifies contextual feature selection with adaptive longitudinal data collection within a single end-to-end trainable model. By leveraging Gumbel-Sigmoid relaxation and straight-through estimation, REACT enables differentiable optimization of discrete acquisition decisions, allowing gradient-based learning to propagate through binary acquisition masks. This approach effectively balances predictive performance against feature acquisition costs. Extensive experiments on multiple real-world health and behavioral longitudinal datasets demonstrate that REACT consistently outperforms existing baselines while achieving superior prediction accuracy at lower total acquisition costs.

Technology Category

• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Search and Optimization: Learning to Search
• Machine Learning: Feature Construction/Reformulation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

In many biomedical applications, measurements are not freely available at inference time: each laboratory test, imaging modality, or assessment incurs financial cost, time burden, or patient risk. Longitudinal active feature acquisition (LAFA) seeks to optimize predictive performance under such constraints by adaptively selecting measurements over time, yet the problem remains inherently challenging due to temporally coupled decisions (missed early measurements cannot be revisited, and acquisition choices influence all downstream predictions). Moreover, real-world clinical workflows typically begin with an initial onboarding phase, during which relatively stable contextual descriptors (e.g., demographics or baseline characteristics) are collected once and subsequently condition longitudinal decision-making. Despite its practical importance, the efficient selection of onboarding context has not been studied jointly with temporally adaptive acquisition. We therefore propose REACT (Relaxed Efficient Acquisition of Context and Temporal features), an end-to-end differentiable framework that simultaneously optimizes (i) selection of onboarding contextual descriptors and (ii) adaptive feature--time acquisition plans for longitudinal measurements under cost constraints. REACT employs a Gumbel--Sigmoid relaxation with straight-through estimation to enable gradient-based optimization over discrete acquisition masks, allowing direct backpropagation from prediction loss and acquisition cost. Across real-world longitudinal health and behavioral datasets, REACT achieves improved predictive performance at lower acquisition costs compared to existing longitudinal acquisition baselines, demonstrating the benefit of modeling onboarding and temporally coupled acquisition within a unified optimization framework.

Problem

• longitudinal active feature acquisition
• contextual descriptors
• temporal features
• cost-constrained prediction
• onboarding phase

Innovation

• longitudinal active feature acquisition
• contextual feature selection
• differentiable optimization
• Gumbel-Sigmoid relaxation
• cost-constrained prediction