NSF supports new AI Institute for Agent-based Cyber Threat Intelligence and OperatioN

Earlier in May, the US National Science Foundation announced an investment of $140 million to launch seven new Artificial Intelligence Research Institutes across the country. One of the seven new institutes will be established at the University of California, Santa Barbara, and includes 11 educational institutions – of which the University of Illinois Urbana-Champaign is one.

Named the NSF AI Institute for Agent-based Cyber Threat Intelligence and OperatioN (ACTION), this cutting-edge research center aims to revolutionize the way mission-critical systems are protected against sophisticated cyber threats.

The institute is set to begin operations on June 1, and the collaborative work will focus on developing intelligent agents that use complex knowledge representation, logic reasoning, and learning to identify and respond to breaches in a timely and scalable fashion. By integrating learning and reasoning, human-agent and agent-agent interaction, and strategic gaming and tactical planning, the institute aims to provide fundamental innovations in several AI domains.

Illinois Computer Science involvement comes from professors Bo Li and Gang Wang.

Li focuses her research on trustworthy machine learning, with an emphasis on robustness, privacy, generalization, and their interconnections. Her group aims to provide trustworthy machine learning systems with guarantees.

In addition to her appointment in CS, she’s on the advisory board of the Center for Artificial Intelligence Innovation (CAII) at Illinois, and she is a member of the Information Trust Institute (ITI). Li also holds affiliations with Electrical and Computer Engineering (ECE), Security and Privacy Research at Illinois (SPRAI), as well as several research centers to broaden research collaboration by bridging different communities.

As a PI for the NSF ACTION Institute, Li said her work will address research thrusts related to the AI stack, such as knowledge-enabled logic reasoning and multi-agent collaboration, and the AI for cybersecurity thrusts, such as cyber vulnerability assessment, detection, attribution, and recovery.

“In the past, our group has developed the data-driven learning with knowledge-enabled logical reasoning framework for different tasks such as image classification, NLP information extraction, and generative models. It is exciting to further explore and deploy the learning-reasoning framework to cybersecurity tasks in practice and provide security and robustness guarantees,” Li said. “The most impactful potential outcomes from the ACTION Institute include a range of fundamental algorithms for AI and cybersecurity, and large-scale AI-enabled systems for cybersecurity tasks with formal security guarantees, which are realized by not only purely data-driven models, but also logic reasoning based on domain knowledge, weak human supervision, and instructions. Such security protection and guarantees will hold against unforeseen attacks, as well.”

Wang serves as a co-PI with Li, and he holds affiliations with ECE at Illinois (by courtesy), SPRAI, and Informatics Programs of School of Information Sciences.

His research that bridges systems work with security and privacy, aims to make it easier to build explainable and robust data-driven solutions to safeguard Internet systems and augment human’s ability to perform security-related tasks.

“With the NSF ACTION institute, I will be primarily working on the research thrusts that seek to enhance the cyber-defense life cycle with the proposed AI-stack,” Wang said. “This involves incorporating AI reasoning and knowledge-based inference into the key cyber defense tasks such as vulnerability assessment, threat detection, attribution, and response and recovery. All of these items are well aligned with my current research, and the project would allow me to work closely with a team of excellent security and AI experts.

“The most exciting prospect is the potential breakthroughs in AI learning and reasoning, and transforming the whole cyber-defense lifecycle to better handle the increasingly sophisticated, ever-changing attacks.”

The ACTION Institute is also committed to being a nexus for educational institutions, government agencies, non-profit organizations, and industry. The institute will train the next generation of the cybersecurity workforce and work towards transitioning novel AI-powered technology into real-world security operations.

“AI is used routinely now, for things like malware analysis to identify malicious documents and malicious webpages,” said UC Santa Barbara computer science professor and cybersecurity expert Giovanni Vigna, who will head this $20 million, five-year project. “What we don’t have are entities that are capable of reasoning. This is an opportunity to bring artificial intelligence and security together in a novel way.”

Per the NSF, this investment is "part of a broader effort across the federal government to advance a cohesive approach to AI-related opportunities and risks."

The announcement came as a number of A.I. chief executives gathered at the White House to discuss recent concerns about their technology's impact on modern society. The new NSF AI Institutes will address some of these concerns as they advance research on ethical and trustworthy AI. Each institute will be an interdisciplinary effort with a unique focus and will draw from a collaborative network of partnering campuses and institutions.

Read the original announcement and an additional story from the University of California, Santa Barbara.