Adversarial Reasoning for Intelligent Information

Abstract

This proposal focuses on game theoretic models of societies and cultures and how they have survived by developing various mechanisms such as social norms and social institutions. These mechanisms have been necessary, since many social interactions entail deception, made possible by the inherent nature of information asymmetry and misaligned utilities, which dictate the strategic interactions in the dynamics of the game. As these conventional structures have begun to be invaded by the ubiquitous Internet technologies which disrupt traditional institutions, exacerbate information asymmetry and degrade trust and respect for privacy, it becomes important to introduce additional mechanisms to guarantee the society-wide liveness and safety properties. The focus of this project is on two kinds of non-strategic agents, e.g., recommenders and verifiers, in order to achieve improved social welfare. Thus, the proposed project only addresses one aspect of a rather ambitious program: Study of how emergent distributed systems with (cryptographically) anonymized agents can be augmented with verification (i.e., blockchain) technologies in order to impact social welfare in multiple and contemporary cyber-social domains; in particular, it seeks to examine the general conditions under which advances to social welfare can be predicted. This approach is strongly rooted in mathematical, scientific and data analysis methodology with critical focus on causality and game theoretic tools. The project is structured around four aims: Aim 1: Sandboxing to develop an expressive dynamic and multi agent game system, instantiated thus far for project management (e.g., GATI: Government as the Integrator, or cancer ICO). Aim 2: Implementing Recommenders and Verifiers who learn. Aim 3: Model Checking the System Behavior and Causality Analysis. Aim 4: Stress Testing (against adversarial Black Swan events).

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 14, 2019
Source ID
W911NF1810427

Entities

People

  • Bhubaneswar Mishra

Organizations

  • Army Contracting Command
  • New York University
  • United States Army

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Game Theory.
  • Systems Analysis and Design

Technology Areas

  • Cyber