Interactive Reasoning and Games with Awareness of Unawareness (Research Topic: Information Sciences)

Abstract

Unknown unknowns are pervasive in defense. The limited situational awareness of a Soldier in a battlefield is likely to imply unknown unknowns. The covert state of mind of an adversary poses unknown unknowns to intelligence officers. A procurement officer who is tasked with the development of a portfolio of novel defense technologies is facing unknown unknowns about capabilities and cost driving factors. These examples are not just situations under plain uncertainty, in which the decision maker can envision all contingencies and judges their likelihood probabilistically. They are also not situations under ambiguity or Knightian uncertainty, in which the decision maker lacks the confidence to judge the likelihood of contingencies probabilistically but still can envision all relevant contingencies. Rather, these are situations, in which the decision maker lacks comprehension or even conception of some of the contingencies. This is unawareness.There is a recent growing interdisciplinary literature on the logic, decision theory, and game theory with unawareness and first applications to economics, contract theory, finance, business strategy, conflict resolution, political science, and strategic network formation. However, one foundational question that is of importance to many potential applications of unawareness remained open: How to model awareness of unawareness in a tractable way? In models of unawareness, if a decision maker is unaware of an event, then she is unaware that she is unaware of that event. Yet, we often realize that there could be something that we are unaware of. This refers to awareness of unawareness. This project develops the logic and game theory with awareness of unawareness. In the first part of the project, we develop multi-agent quantified unawareness structures that allow for reasoning about knowledge of unawareness. The second part of the project extends this to quantified unawareness type-spaces that allow for reasoning about probabilistic beliefs of unawareness. These type-spaces are applied in the third part of the project, which develops Bayesian games with awareness of unawareness of actions, players, and payoƀ-relevant events. This class of games is then applied to the development of mechanism design under awareness of unawareness. Potential applications relevant to the mission of the Army comprise of mechanisms for the procurement and the management of portfolios of novel weapons, security games, and support systems for decisions under awareness of unawareness.

Document Details

Document Type
DoD Grant Award
Publication Date
Sep 28, 2022
Source ID
W911NF2210282

Entities

People

  • Burkhard Schipper

Organizations

  • Army Contracting Command
  • United States Army
  • University of California, Davis

Tags

Readers

  • Artificial Intelligence
  • Canine Service Warrior Training Program for Wounded Warriors in the Veterinary Industry, Supported by Donors.
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • Space