On Myopia in Rationally Inattentive Dynamic Information Acquisition

Arefanjazi, Hadis; Akyol, Emrah

Permalink

https://hdl.handle.net/2142/130301 Copy

Description

Title

On Myopia in Rationally Inattentive Dynamic Information Acquisition

Author(s)

• Arefanjazi, Hadis
• Akyol, Emrah

Issue Date

2025-09-17

Keyword(s)

• Human–AI collaboration
• Dynamic information acquisition
• Rational inattention
• Majorization theory
• Matrix norms

Abstract

In many human–AI collaborative contexts, a decision maker must sequentially allocate limited attention to acquire information, culminating in a single decision at an exogenously determined time. The overall objective is to minimize a given function of the estimation error covariance at that final decision time. While the decision maker knows the current objective function, future objectives remain unspecified. Extending recent insights from economics on similar dynamic information acquisition problems, we use majorization theory to analyze conditions under which the optimal attention allocation exhibits monotonicity across time stages and remains time-consistent. Specifically, we characterize the conditions on the objective and attention functions that guarantee that a strategy chosen without full knowledge of future objectives remains optimal. Our results highlight the robustness of such allocation policies, offering guidelines for human–AI collaborations where partial or evolving knowledge of performance measures often characterizes the decision-making environment.

Publisher

Allerton Conference on Communication, Control, and Computing

Series/Report Name or Number

2025 61st Allerton Conference on Communication, Control, and Computing Proceedings

ISSN

2836-4503

Type of Resource

Text

Genre of Resource

Conference Paper/Presentation

Language

eng

Handle URL

https://hdl.handle.net/2142/130301&&

Copyright and License Information

Copyright 2025 is held by Hadis Arefanjazi and Emrah Akyol.

Owning Collections