Secure Distributed Computation and Learning Networks


The central goal of this project is to advance our understanding of what it takes to make various large-scale decision-making, computation, optimization and machine learning algorithms 'secure' against network and/or component failures and adversary attacks. The class of algorithms to be considered are those intended to enable the functioning of distributed networks such as sensor networks, computer networks, and multi-robot teams. The algorithms devised will be demonstrated and validated in extensive simulations and possibly a multi-robot testbed. The project aims to develop a framework and computational models for teams of heterogeneous agents to reliably exploit information from various sources in uncertain and potentially unfriendly environments.

Issues Involved or Addressed

In many cyber-physical applications, large volumes of heterogeneous streaming data are needed to be collected by a team of autonomous agents which then collaboratively explore a complex and cluttered environment to accomplish various types of missions. In such a distributed system, groups of agents perform a multitude of operations, including decision making, computation and estimation, optimization and learning with streaming data. To successfully and securely perform these operations in uncertain and unfriendly environments, novel concepts and methodologies are needed to craft and analyze secure algorithms capable of reliably delivering information and robustly performing desired operations.

Methods and Technologies

  • Cyber security
  • machine learning
  • optimization
  • privacy preserving
  • distributed computing
  • multi-agent systems
  • sensor networks
  • multi-robot systems

Academic Majors of Interest

  • Electrical Engineering
  • Computer Science
  • Computer Engineering
  • Mechanical Engineering
  • Applied Mathematics and Statistics
  • Mathematics
  • Civil Engineering
  • Information Systems

Preferred Interests and Preparation

Team Advisors