Top 8 Leading-Edge MURI Projects for FY 2024: Advancing Cybersecurity

In today’s rapidly advancing world, the challenges we face in science and engineering, particularly in cybersecurity, are increasingly complex, often requiring a multidisciplinary approach to find innovative solutions.

Dr. Bindu Nair, director of the Basic Research Office in the Office of the Under Secretary of Defense for Research and Engineering, highlights this approach, stating: “The MURI program acknowledges these complexities by supporting teams whose members have diverse sets of expertise as well as creative scientific approaches to tackling problems.”

By fostering collaboration across various disciplines, the MURI program addresses these challenges head-on. Before diving into the top eight groundbreaking MURI projects for Fiscal Year 2024 – each focused on innovative cybersecurity solutions – let’s first understand how the program supports both scientific progress and national defense.

Addressing Cybersecurity Threats: How the MURI Program Enhances National Defense

Cybersecurity threats, such as data breaches, ransomware attacks, and state-sponsored hacking, are growing and pose serious risks to individuals, businesses, and governments.

In 2023, according to the Deloitte Cybersecurity Threat Trends Report, 66% of organizations experienced ransomware attacks, 44.7% of data breaches involved the misuse of valid credentials, and there was a staggering 400% increase in IoT malware attacks, with the manufacturing industry being the most targeted.

These threats can cause financial losses, expose sensitive information, and disrupt vital services.  

To address these dangers, the Department of Defense (DoD) plays a key role in improving national security. The DoD invests in research and funds projects like the Multidisciplinary University Research Initiative (MURI). These efforts bring together experts from universities, businesses, and government agencies to create stronger defense strategies.

The MURI program tackles complex scientific and engineering problems that can’t be solved by one field alone. Over the past 25 years, it has funded over 600 projects, helping to make progress in many areas.

The program stands out for its approach of bringing together experts from different fields to work on difficult problems. This helps increase scientific knowledge and strengthen national security.

In 2024, MURI is supporting projects in areas like cybersecurity, machine learning, quantum systems, and materials science. The Department of Defense announced $221 million in funding for basic defense research through MURI. Each project typically receives about $7.5 million over five years.

The MURI program is important because it combines knowledge from different areas to solve tough challenges, speeds up research for faster breakthroughs, and develops new technologies that protect national security.

Through fostering collaboration and offering substantial funding, the DoD ensures that the U.S. remains at the forefront of technological advancements and defense.

Top 8 MURI-Selected Projects Advancing National Security and Technology

Let’s take a closer look at the Top 8 MURI-Selected Projects in 2024 that are driving advancements in cybersecurity field.

MIT’s Research on Real-Time Interventions for Large Networked Systems: Evaluation and Optimization

The project “Evaluating, Predicting, Optimizing, and Monitoring Hypothetical Interventions in Large Networked Systems” belongs to the broader research topic “Interventions in Large Networked Systems: Prediction, Monitoring and Evaluation.”

Spearheaded by Caroline Uhler at the Massachusetts Institute of Technology (MIT), with collaboration from Harvard University, Princeton University, and Leland Stanford Junior University, this research aims to develop sophisticated methods for predicting and optimizing interventions in complex networks to enhance their stability and performance.

Uhler and collaborators will use the award to improve intervention strategies in complex systems, enhancing decision-making across biomedical, engineering, and societal fields.

Network theory helps understand the complex relationships and dynamics in large systems, while machine learning analyzes data and makes accurate predictions. This approach uses the strengths of both fields to improve network management.

The potential impact of this research is substantial. Critical infrastructure networks, such as power grids and communication systems, are increasingly complex and interconnected, making them vulnerable to disruptions. By developing tools to predict and mitigate these disruptions, the project aims to improve the resilience and efficiency of these essential systems.

For instance, in the context of a power grid, the ability to predict potential failures and optimize interventions could prevent widespread outages and ensure a stable supply of electricity.

Moreover, the real-time aspect of the predictive models means that interventions can be dynamically adjusted as conditions change, providing a more responsive and adaptive approach to managing complex networks.

This capability is particularly crucial in scenarios where rapid decision-making is required to prevent cascading failures or to respond to emerging threats.

Improving Decision-Making in Autonomous Systems: Columbia University’s MURI Project

The project titled “Algorithms, Learning, and Game Theory: The Foundations of Multi-Agent Systems” falls under the broader research area of “Theory and Algorithms for Learning and Decision-Making in Multi-Agent Systems.”

Led by Christos Papadimitriou at Columbia University, in collaboration with MIT, the University of Maryland, College Park, Carnegie Mellon University, and the University of California, San Diego, this research aims to create foundational algorithms and game theory models to improve decision-making in systems involving multiple autonomous agents.

This initiative is part of the Fiscal Year 2024 MURI program, with each team typically receiving an average award of $7.5 million over five years. The project combines game theory with machine learning to develop robust decision-making frameworks for multi-agent systems.

Game theory provides a structured approach to understanding the interactions between rational agents, while machine learning helps analyze data and optimize decision-making processes.

The research has the potential to make an impact on fields that rely on autonomous systems, such as defense, logistics, and other critical applications. These systems often require multiple agents to collaborate and make decisions in complex settings. 

For example, in defense, improved coordination between autonomous drones could enhance missions such as surveillance and reconnaissance. In logistics, better decision-making models could streamline the scheduling and routing of delivery vehicles, leading to cost savings and more efficient service.

Multi-Agent Systems with New Game Theory: A Look at the ‘New Game Theory for New Agents’ Project

Let’s take a look at another project under the topic “Theory and Algorithms for Learning and Decision-Making in Multi-Agent Systems.” It is titled “New Game Theory for New Agents: Foundations and Learning Algorithms for Decision-Making in Mixed-Agent Systems.

This project, headed by Vijay Subramanian at the University of Michigan, focuses on creating new game theory models and learning algorithms designed to enhance decision-making in systems where diverse types of agents collaborate.

“There are lots of different agents that are interacting, including the usual players––humans––which could be big entities, like corporations, governments, or other institutions,” explained Vijay Subramanian, professor of Electrical and Computer Engineering and project director. “But in today’s world, we have these new AI agents as well. What we want to understand is: how do these computational agents interact?”

The project is being conducted in collaboration with Yale University, Toyota Technological Institute of Chicago, the University of Southern California, MIT, Harvard University, California Institute of Technology, and Cornell University.

This project, part of the Fiscal Year 2024 MURI program, introduces new game theory ideas made for systems with different types of agents. It aims to make these systems more adaptable and stronger, focusing on how diverse agents work together and make decisions in complex situations.

This research could have a big impact. Multi-agent systems are used in many areas, like cybersecurity, where different types of agents need to work together and make decisions.

By creating better game theory models and learning algorithms, the project aims to improve the security and efficiency of these systems. For example, in cybersecurity, the improved decision-making could help agents work together to detect and respond to threats more effectively.

“Our goal is to transcend that and develop new theory that can address this mixture of autonomous, semi-autonomous, algorithmic, and human agents,” said Subramanian.

By predicting the outcomes of interactions involving AI agents, the research team can design environments and tasks that are executed more effectively and with greater precision.

A practical example of where this analysis would be beneficial is in disaster recovery operations, such as those following an earthquake or airstrike. In such situations, humans and robots may collaborate to clear debris and provide medical assistance to injured individuals.

AI-Guided Self-Organization at Yale University: Shaping Complex Nonlinear Dynamics for Enhanced System Stability and Security

“AI-Guided Self-Organization: Tailoring Disorder to Shape Complex Nonlinear Dynamics” is also a notable project of the Fiscal Year 2024 MURI program, led by Hui Cao at Yale University.

This project, part of the broader research area of “Complexity Science Disorder-Promoted Synchronization,” involves collaboration with the University of Michigan, University of Maryland, College Park, Arizona State University, Virginia Tech, and Wesleyan University.

The project aims to use artificial intelligence to solve the challenge of getting quantum photonic sensors to work together. The goal is to improve their power, emission, and signal coherence.

“We are willing to try new techniques. But it is the charge of the MURI awards to push fundamental research even if it is unsuccessful because we will still be learning new things.” said Vassilios Kovanis, a quantum systems research professor within the Virginia Tech National Security Institute, who is on one of the multi-university teams awarded.

“The MURI award is based on the idea of encouraging collaboration between peer institutions,” Kovanis said. “There is so much more we can learn when we are able to work together with our peers at different universities who may have different resources than us, but also different perspectives.”

The potential impact of this research is far-reaching. Complex systems, like those encountered in cybersecurity, are often dynamic and unpredictable. The project aims to develop methods for directing the self-organization of these systems, which could enhance their stability and performance.

This five-year project, funded by the Office of Naval Research, will focus on basic research to discover new physical principles and AI methods that can universally control and regulate self-organization. The team will also emphasize training the next generation of defense scientists and engineers.

Key goals include achieving reconfigurable phase-locking in 2D semiconductor laser arrays, scalable power mode-locking in fiber lasers, and shaping emissions from arrays of Josephson junctions, a type of quantum oscillator.

Enhancing Autonomous Systems with REPRISM: Cognitive Science Meets Problem-Solving

Another exciting project under the Fiscal Year 2024 MURI program is “REPRISM: Flexible Embodied Problem-Solving by Manipulating the Representational Prism,” which falls under the broader research area of “Cognitive and Neuroscience-Inspired Problem-Solving for Autonomous Systems in Physical Environments.”

This research, led by George Konidaris at Brown University in collaboration with MIT, focuses on developing cognitive and neuroscience-based approaches to enhance problem-solving abilities in autonomous systems.

This project has received substantial funding, with each team getting ample support. It aims to use ideas from cognitive science and neuroscience to make autonomous systems smarter and more adaptable. By mimicking how the human brain solves problems, the goal is to create systems that can handle complex situations with flexibility and efficiency.

The potential applications and impact of this research are far-reaching. In domains like cybersecurity, autonomous systems must often deal with evolving and unpredictable challenges.

This research aims to equip these systems with enhanced problem-solving capabilities, allowing them to adapt quickly to new threats and safeguard critical data and infrastructure.

Neurocognitive Approaches to Overcoming Failures in Autonomous Systems at Carnegie Mellon University

The project titled “Overcoming Unexpected Failures Using Neurocognitive Multi-Abstraction Active Exploration” is also part of the broader research theme “Cognitive and Neuroscience-Inspired Problem-Solving for Autonomous Systems in Physical Environments.”

The research, which applies principles from cognitive science and neuroscience, is led by David Held at Carnegie Mellon University, with collaboration from the University of Massachusetts, Amherst, MIT, Princeton University, Stanford University, and the University of Washington.

It focuses on developing techniques that help autonomous systems overcome unexpected failures.

By integrating neurocognitive principles with active exploration techniques, this approach strives to enhance the adaptability and resilience of autonomous systems. The goal is to build systems capable of managing unforeseen challenges and responding to evolving threats, drawing on cognitive science and neuroscience insights to improve their decision-making.

The impact of this research could be great, particularly for autonomous systems in cybersecurity, where unpredictable obstacles and emerging risks are frequent. The project aims to develop innovative methods that enable these systems to recover quickly from failures and adapt to new circumstances.

For example, in cybersecurity, applying neurocognitive principles could empower autonomous systems to tackle sophisticated cyber threats, ensuring the protection of critical data and infrastructure.

Understanding Belief-Narrative Resonance in the Age of Generative AI: The BRAIN Project

Another exciting project under the broader research topic “Modeling and Measuring Multilevel Resonance” is titled “BRAIN (Belief Resonance and AI Narratives): Understanding Belief-Narrative Resonance in the Era of Generative AI.”

Led by Yong-Yeol Ahn at Indiana University, in collaboration with New York University and the University of California, Berkeley, this project explores how beliefs and narratives influence behavior, particularly in the context of generative AI.

“The deluge of misinformation and radicalizing messages poses significant societal threat,” said lead investigator Yong-Yeol Ahn.

“Now, with AI, you’re introducing the potential ability to mine data about individual people and quickly generate targeted messages that appeal to them — applying big data to individuals — which could cause even greater disruptions than we’ve already experienced.” he added.

As part of the Fiscal Year 2024 MURI program, this initiative blends artificial intelligence with social science to study the effects of narratives on societal behavior – a novel approach that could bring new insights into the realm of information security.

Ahn explained that the project will explore the concept of “resonance” and its impact on how people respond to certain messages. Resonance suggests that individuals are more influenced by content that aligns with their emotions or pre-existing beliefs, such as political views, religious beliefs, or cultural norms.

Resonance can be used to either bridge divides between groups or deepen polarization. With AI’s ability to quickly generate personalized content through text, images, or videos, the potential to amplify the effects of these messages – positively or negatively – has increased.

The team will use AI technology to enhance their research, employing “model agents” – virtual personas that share information and respond to messages in simulations. This will enable the team to more accurately study the flow of information.

Understanding the spread of misinformation and its influence on behavior is critical for developing more effective cybersecurity strategies.

Ahn also noted that the IU-led team’s work will mark a shift from previous efforts to model belief systems by simulating people’s opinions. The project will incorporate “a complex network of interacting beliefs and concepts, integrated with social contagion theory,” aiming to create “a holistic, dynamic model of multi-level belief resonance.”

The goal is to develop a system that better reflects the complexities of real life, where individuals’ opinions are shaped not just by political affiliation but by a blend of various belief systems and social dynamics.

Harnessing Tensor Networks for High-Dimensional Kinetic Simulations

The project “Tensor Approaches for Simulating Kinetic Systems” falls under the research topic “Tensor Networks and Low-Rank Methods for High-Dimensional Computing.” It is led by Jingmei Qiu at the University of Delaware.

With collaborations from Michigan State University, the University of Washington, the University of Illinois Urbana-Champaign, Stanford University, and the University of California, Santa Cruz, this research focuses on utilizing tensor network methods to simulate complex kinetic systems in high-dimensional spaces.

Challenges like kinetics and particle distributions in nonequilibrium plasmas involve solving partial differential equations in high-dimensional spaces, which traditional simulation methods struggle to address.

The research team aims to tackle these issues using the tensor network approximation, which could significantly lower the computational cost of studying high-dimensional systems. While the focus is on nonequilibrium plasmas, the developed tools will also be applicable to a range of other problems relevant to the Department of Defense.

The use of tensor networks in high-dimensional computing represents a cutting-edge approach to improve the efficiency and accuracy of simulations. The project aims to tackle the challenges associated with simulating kinetic systems, which often involve high-dimensional data and complex interactions.

What’s more, advanced computational tools are crucial for modeling complex systems and identifying emerging threats. By developing tensor network methods, this project aims to improve the ability to simulate complex systems, thereby enhancing cybersecurity strategies.

For example, better modeling of network traffic and threat behaviors could lead to more effective detection and prevention of cyber-attacks, ensuring the security and integrity of critical infrastructures.