April 19, 2019

Integrating acquisition and analysis in AI systems

UCR engineering professors receive $1.2M Navy grant to make artificial intelligent (AI) systems smarter

Author: Katharine Hall
April 19, 2019

BCOE ONR grant team
Amit Roy-Chowdhury, Fabio Pasqualetti, Salman Asif and Konstantinos Karydis

AI systems have largely focused on analyzing data that has already been collected. That may be changing soon thanks to a $1.2M grant from the Office of Naval Research’s Science of Artificial Intelligence Program awarded to four UCR engineering professors. Building upon existing low-level feature extraction and classification methods, researchers will focus on the high-level machine learning and decision making process to automatically integrate data analysis and data acquisition phases without human interaction.

Current AI systems can process very large amounts of information to aid in decision making but research by Amit Roy-Chowdhury, Konstantinos Karydis, and Salman Asif, electrical and computer engineering professors, along with Fabio Pasqualetti, mechanical engineering professor, proposes a new way AI can boost efficiency. Using recent advances in computer vision, machine learning, signal processing, robotics, and controls, researchers will explore AI’s capability of both interpreting large amounts of data and acquiring it smartly so the systems can expedite making informed decisions.

For field operations ranging from national security to disaster response, there is a need for understanding environments from large heterogeneous data collected by a team of autonomous agents. Algorithms can be trained to complete narrow, specific, anticipated tasks with describable and predictable outcomes. However, most current methods require significant human supervision. The proposed research will enable continuous and efficient learning of models that represent the data across sensors and modalities without the need for strong supervision. This enables those learning models to make decisions about the most-task relevant portions of the data, detect any adversarial attacks, and collaboratively plan to acquire additional data to maximize decision making accuracy. This collaboration addresses some of the fundamental limitations of current autonomous systems by requiring only limited supervision, having the ability to identify relevant and reliable portions of large data volumes, and integrating the analysis and acquisition phases.

Artificial intelligence is increasingly fast, agile, and seemingly low-cost. In the age of digital supremacy, error reduction in these intelligent systems is vital. Using AI to not only analyze data but also acquire it can lead to smarter technologies in our digital world.