Biography

Hi! I'm Valerie Liang, a student at Johns Hopkins University studying Computer Science, Applied Mathematics and Statistics, and Cognitive Science. I'm fascinated by how machine learning can bridge disciplines—whether that's helping robots “understand” humans better, improving medical analyses, or making engineering systems more adaptive. For me, the true power of machine learning isn't in its theoretical complexity, but in its ability to create deceptively simple solutions that have a profound impact on the world. Some of my favorite projects have involved reinforcement learning for collaborative systems, computer vision for accessibility, and data-driven modeling in cognitive science. What excites me most about research is how even a small insight or improvement can ripple outward and create real impact. I also enjoy software engineering and building tools that make technology more approachable for everyone, because I believe innovation should be universally accessible, not locked behind commercialization. Outside of research, I'm part of a couple of performance groups, including lion dance and fire spinning. I like anything flashy and high-energy: it's a fun way to stay creative and connect with people outside the lab.

Academic Status

Undergraduate Student - 3rd

Research Area/Department

Applied Mathematics; Computer Science

Major/Specialty

Computer Science, Applied Mathematics and Statistics, and Cognitive Science Focus on machine learning and computer vision - ML Expertise: Reinforcement Learning, Deep Learning, Computer Vision, Neural Networks, Data Analysis, Optimization

Degrees Earned or in Progress

B.S. Computer Science, B.S Applied Mathematics and Statistics, B.A. Cognitive Science; August 2023 - May 2027

Academic Preparation

Data Structures, Intermediate Programming, Mathematical Foundations for Computer Science, Linear Algebra, Calculus I/II/III, Computer Ethics, Full-Stack Javascript, Professional Writing and Communication, Computer Systems Fundamentals, Honors Probability, Mathematical Statistics, Mathematical Image Analysis, Algorithms, Artificial Intelligence, Computer Vision, Optimization

Research/Publications

Machine Learning Research Intern during a National Science Foundation-sponsored Summer REU: May 2025 - Aug. 2025, Complex Systems Monitoring, Optimization, and Stability Lab - University of Tennessee, Knoxville; I'm currently continuing this research independently for college credit while our paper is under review - Developed deep reinforcement learning algorithms (PyTorch, TensorFlow) for decentralized UAV & ground vehicle control. - Designed rendezvous location mapping frameworks for truck-drone delivery systems, achieving 48% efficiency gains in logistics simulations. - Implemented multi-agent learning architectures with state–action exchange to optimize cooperative performance. Reference: Sabrullah Deniz (sdeniz@vols.utk.edu) Publications: Liang, V., Deniz, S., & Chakraborty, S. (2025). Decentralized Multi-Agent Reinforcement Learning for Hybrid Truck-Drone Last-Mile Delivery. Transportation Research Board Annual Meeting. In review. Undergraduate Research Assistant: Oct. 2024 - Present, Kennedy Krieger Institute - Optimized MATLAB pipelines for large-scale data cleaning, preprocessing, and analysis of neurocognitive datasets. - Applied computational methods to quantify experimental patterns and validate hypotheses. - Delivered insights from literature reviews to guide research decisions and methodologies in ongoing projects. References: Megan Markiewicz (markiewicz@kennedykrieger.org)

Research/Academic Interests

My research and academic interests lie at the intersection of computer science, applied mathematics, and cognitive science, with a focus on developing machine learning and optimization methods that address real-world challenges. I am particularly interested in reinforcement learning, computer vision, and computational modeling, and how these techniques can be applied to domains such as robotics, healthcare, and human-centered systems.

Computational and Data Science Areas

Applied Computer Science; Applied Mathematics; Artificial Intelligence and Intelligent Systems; Computer Science; Statistics and Probability; Visualization and Human-Computer Systems

Motivation

I’ve always found technology to be most valuable when it has a widespread effect, not only because it brings convenience to humans’ everyday lives, but because it facilitates more understanding, accessibility, and shared curiosity. While I think researching machine learning and artificial intelligence theory has its own merit, the interdisciplinary applications of ML/AI research are where this powerful technology can be used to facilitate discovery in other fields. I see AI not just as a tool to automate labor, but as a way to amplify human insight—contextualizing and making sense of complex systems so we can solve problems at a scale and depth that was once unimaginable. I appreciate the mission of the Sustainable Research Pathways program and the Sustainable Horizons Institute because I also believe that true innovation is unlocked by a thriving, inclusive community of different perspectives and ideas. SRP’s approach of inviting researchers from all backgrounds promotes an environment that allows for building connections and ensuring widespread discovery, leading to more innovative and robust scientific outcomes for everyone. This past summer, I contributed to a National Science Foundation-funded research internship. My work focused on decentralized reinforcement learning for cooperative truck-drone delivery systems, particularly, multi-agent coordination between ground vehicles and UAVs. My solution was to design a novel model that achieved a 48% reduction in delivery time in real-world simulations. The project's interdisciplinary nature—spanning aerospace vehicle operations to civil engineering road logistics—required me to quickly master domain-specific concepts like drone flight dynamics, which were outside my computer science background. This fostered a collaborative, two-way knowledge exchange; I applied my expertise to advance the project's methodology while instructing peers on reinforcement learning principles and our multi-agent architecture, while they taught me about concepts like Momentum Theory and taught me how to actually fly drones. My commitment to this research extended well beyond the program's conclusion, culminating in its recent acceptance for presentation at the Transportation Research Board Annual Meeting. Successfully elevating my summer's work to a peer-reviewed publication has reinforced my commitment to pursuing technically rigorous research with demonstrable real-world impact. My research trajectory has given me a deeply practical perspective on applied AI. In my current role at the Kennedy Krieger Institute’s Center for Neurodevelopmental and Imaging Research, I specialize in the computational foundations of neurocognitive studies. While my primary responsibility was data preprocessing, I identified a critical bottleneck: researchers were spending excessive time on manual, tedious data quantification tasks. To solve this, I proactively designed and implemented optimized, automated pipelines that extracted key feature data from participant samples, freeing researchers from cumbersome pre-processing, reclaiming their time for higher-value analysis, and accelerating the pace of our research. Learning from professionals like Dr. Musad Haque, who works at APL, gave me a practical view of the AI field I want to enter. His lectures moved beyond theory, detailing the challenges of developing robust multi-agent systems for space robotics, such as managing communication latency and ensuring autonomous decision-making in unpredictable environments. This was a compelling, real-world illustration of the work I find most meaningful: solving complex, large-scale problems through cooperative and autonomous systems. Seeing how these principles are applied to projects like satellite servicing or planetary exploration reinforced my ambition to contribute to similarly impactful work. Similarly, my visit to the Oak Ridge National Laboratory provided a valuable, behind-the-scenes look at a major research institution. Engaging with researchers like Dr. Peter Fuhr, who discussed his work on sensor networks for critical infrastructure, and touring facilities like the Oak Ridge Leadership Computing Facility (OLCF) allowed me to see the collaborative dynamic firsthand. I witnessed how teams bridge disciplines, combining domain expertise in material science, engineering, and computer science to achieve breakthroughs. This experience confirmed that an environment that blends deep, specialized knowledge with a mission-driven purpose is precisely where I can contribute most effectively. Ultimately, my aim is a research career where I can tackle complex problems alongside other dedicated scientists. The SRP program represents the ideal environment for this, offering a direct pathway to contribute to meaningful work while learning the collaborative practices that drive innovation at a high level. There is no substitute for firsthand experience, and the opportunity to work directly with passionate experts on consequential problems is exactly the kind of experience I am seeking. I am eager to contribute to a team while gaining a practical understanding of how to build a meaningful career at the intersection of AI and public impact.

Lightning Talk Title

AI That Works: Interpretable Models for Complex Systems

Keywords (Maximum 20 words)

AI; machine learning; multi-agent systems; computer vision; robotics; reinforcement learning; dynamic environments; feature extraction; practical AI; AI architectures; interpretable AI