Department Seminars

Please note this seminar will take place in 228 Malone Hall.

Zoom link »

Computer Science Seminar Series

“Experimental Design for Data-Efficient Scientific Discovery”

Abstract: Many scientific and engineering problems are constrained not by model expressiveness, but by the high cost of data acquisition, requiring learning systems to make sequential decisions under uncertainty. In this talk, Quan Nguyen presents a research program on data-efficient learning and decision-making, grounded in probabilistic modeling and decision theory. He will first discuss algorithmic approaches to non-myopic search problems, illustrating how explicit lookahead and long-horizon reasoning improve performance compared to greedy baselines. He will then turn to active learning and information valuation, which are relevant in many scientific discovery settings where the aim is not purely optimization or search, but efficient learning of complex systems. Nguyen will introduce an information-theoretic criterion, Vendi Information Gain, that quantifies the value of candidate observations by jointly capturing informativeness and diversity, and provides stable and interpretable guidance for data acquisition. He will present algorithmic designs and empirical results demonstrating improved data efficiency across synthetic benchmarks and real scientific applications, and conclude by outlining a vision for practical, uncertainty-aware experimentation pipelines that integrate sequential decision-making in AI-driven scientific discovery.

Speaker Biography: Quan Nguyen is a postdoctoral researcher at Princeton University whose work focuses on data-efficient machine learning and sequential decision-making. His research develops principled methods for deciding what data to collect next when observations are costly or limited, with real-world applications in scientific discovery and decision support. His doctoral research received the Washington University in St. Louis Turner Dissertation Award for contributions to learning and decision-making under uncertainty.

View the recording »

Computer Science & Biomedical Engineering Seminar Series

February 11, 2026

Abstract: Modern genomics produces massive, high-dimensional datasets, yet extracting reliable biological insight remains challenging. A central challenge—and opportunity—in genomics is that we do not directly receive a sanitized data matrix: Our machine learning pipeline starts upstream, with what samples we choose to collect, how we measure them, and how we transform raw signals before inference. These stages are typically handled in isolation, quietly introducing bias, discarding information, and limiting discovery potential. Tavor Baharav’s research addresses this by treating genomics as an end-to-end system, developing rigorous machine learning methods that account for—and leverage—these upstream choices. In this talk, Baharav will illustrate this approach through his work on reference-free genomic analysis. Alignment of reads to a reference genome, though ubiquitous, fundamentally limits discovery of novel biology that deviates from the reference. To overcome this, Baharav’s team developed SPLASH, a statistical tool that compares raw sequencing reads directly across conditions. SPLASH rediscovers strain-defining mutations in SARS-CoV-2 and identifies previously unannotated tissue-specific transcripts in the octopus genome, enabling discovery without any reference or annotation. Bypassing alignment reshapes the statistical problem: To identify genomic features of interest, his team developed a new statistical test for contingency tables. Aggregating information across the resulting data matrices raised broader methodological and theoretical questions about data integration, leading the team to develop a random matrix theory framework for detecting shared structure across datasets. Together, these results show how rethinking upstream pipeline choices can simultaneously improve biological discovery and yield generalizable statistical insights.

Speaker Biography: Tavor Baharav is a postdoctoral fellow at the Eric and Wendy Schmidt Center at Broad Institute, working with Rafael Irizarry. His research co-designs the machine learning pipeline for computational genomics, jointly optimizing upstream processing stages with downstream inference. Baharav is broadly interested in high-dimensional statistics, adaptive algorithms and statistical machine learning, as well as their application to problems in computational genomics. Prior to his postdoctoral work, he earned his PhD in electrical engineering from Stanford University in 2023 under the guidance of David Tse and Julia Salzman and funded by the NSF Graduate Research Fellowship and the Stanford Graduate Fellowship. Baharav’s research at the intersection of machine learning and genomics has been published in venues ranging from the Conference on Neural Information Processing Systems and the Journal of Machine Learning Research to the Research in Computational Molecular Biology conference and Cell.

View the recording »

Institute for Assured Autonomy & Computer Science Seminar Series

February 9, 2026

Talk 1: “A Perspective on Safety, Risk, and Reproducibility in Reinforcement Learning”

Abstract: Reinforcement learning is a field that has gained salience in recent years, dating back to the 1990s with backgammon-playing bots, and more recently with AlphaGo and training mechanisms for large language models. Reinforcement learning may refer to a problem, a learning paradigm, a collection of techniques, and so on—but irrespective of its designation, it is rife with disparate sources of randomness. We provide some perspective on how communities—spanning operations research, computer science, electrical engineering, and statistics—have tried to quantify and mitigate them. These mitigation strategies include safety, learning with constraints, risk measures, reproducibility, simulator calibration, and data coverage concerns. Time permitting, we’ll cover some emerging techniques to enforce reproducibility constraints into the standard RL training pipeline.

Speaker Biography: Alec Koppel is a senior professional staff member at the Johns Hopkins University Applied Physics Labs within the Artificial Intelligence/Machine Learning Group in the Research and Exploratory Development Mission Area. Previously, he was an AI research lead in the Multiagent Learning and Simulation Group at JPMorganChase Artificial Intelligence Research. He was also a research scientist in optimal sourcing systems within Amazon’s Supply Chain Optimization Technologies team and a research scientist at the U.S. Army Research Laboratory Computational and Information Sciences Directorate from 2017 to 2021. Koppel’s research focuses on optimization and machine learning, spanning applications in autonomous systems/robotics, financial networks, and supply chain optimizations, with particular interests in reinforcement learning/bandits, scalable online Bayesian and nonparametric methods, and online learning and stochastic optimization.

Talk 2: “Reliable Reinforcement Learning for Robot Autonomy”

Abstract: Pratap Tokekar will review recent trends in safe and reliable reinforcement learning for the control of robotic autonomous systems. The Robotics Algorithms & Autonomous Systems Lab designs algorithms and builds systems to enable teams of robots to act as sensing agents, with research at the intersection of theory and systems motivated by real-world applications to environmental monitoring, infrastructure inspection, and precision agriculture.

Speaker Biography: Pratap Tokekar is an associate professor in the Department of Computer Science at the University of Maryland and an Amazon Scholar at Amazon Robotics. Between 2015 and 2019, he was an assistant professor in the Bradley Department of Electrical and Computer Engineering at the Virginia Polytechnic Institute and State University. Previously, Tokekar was a postdoctoral researcher at the University of Pennsylvania General Robotics, Automation, Sensing, & Perception Lab. He obtained his PhD in computer science from the University of Minnesota in 2014 and his bachelor of technology degree in electronics and telecommunication from the College of Engineering, Pune in India in 2008. Tokekar has received a 2022 Amazon Research Award, a 2020 NSF CAREER Award, and a 2016 NSF Computer and Information Science and Engineering Research Initiation Initiative award. He has also served as an associate editor for IEEE Transactions on Robotics and on the editorial boards for the IEEE International Conference on Robotics & Automation and the IEEE/RSJ International Conference on Intelligent Robots and Systems.

View the recording »

Computer Science & Biomedical Engineering Seminar Series

February 5, 2026

Abstract: Understanding the mechanisms linking genetic sequence to cellular function remains a central challenge in biology. Existing computational approaches are often computationally expensive or fail to generalize beyond well-studied model organisms and protein families. Samuel Sledzieski’s work leverages protein language models to connect molecular structure to systems biology at genome scale. He will highlight three applications where protein language modeling unlocks new capabilities: training on molecular dynamics simulations to predict protein conformational dynamics, enabling an analysis of the allosteric behavior of KRAS; building de novo protein-protein interaction networks in non-model organisms, revealing previously uncharacterized proteins involved in coral stress response; and high-throughput screening of massive small molecule libraries for drug-target interactions, identifying novel kinase inhibitors with experimentally validated nanomolar affinity. Sledzieski will discuss how recent advances in contrastive learning, parameter-efficient fine-tuning, and multimodal representation learning address key computational barriers to effective genome-scale modeling. Finally, he proposes a research plan to model heterogeneity in protein structure and molecular interactions across cellular contexts.

Speaker Biography: Samuel Sledzieski is a Flatiron Research Fellow at the Flatiron Institute Center for Computational Biology and a visiting researcher at the Lewis-Sigler Institute for Integrative Genomics at Princeton University. His research uses protein language models to integrate molecular biophysics with systems genomics, with the ultimate goal of mapping the mechanisms of cellular behavior and complex disease. He has developed and released several open-source machine learning models, including D-SCRIPT, ConPLex, and RocketSHP, and has held research positions at Microsoft Research, Cellarity, Serinus Biosciences, and the Centre Scientifique de Monaco. A recipient of the NSF Graduate Research Fellowship, Sledzieski received his PhD (2024) and MS (2021) in computer science from the Massachusetts Institute of Technology, and his BS (2019) from the University of Connecticut.

View the recording »

Computer Science Seminar Series

January 29, 2026

Abstract: Machine learning and AI are not standalone artifacts: They are ecosystems where foundation models are adapted and deployed through layered pipelines spanning developers, platforms, users, and regulators. This talk explores how the structure of these ecosystems shapes the distribution of value and risk, and determines system-level properties like safety and fairness. Benjamin Laufer begins with a game-theoretic model of the interaction between general-purpose producers and domain specialists, using it to examine how regulatory design shapes incentives and equilibrium behaviors. He then connects these formal insights to empirical measurements from 1.86 million open-source AI models, reconstructing lineage networks to quantify how behaviors and failures propagate through fine-tuning. Finally, turning from the descriptive structure of the ecosystem to the design of the algorithms themselves, Laufer describes his work in algorithmic fairness, framing the identification of less discriminatory algorithms as a search problem with provable statistical guarantees. He closes by outlining a forward-looking research agenda aimed at building technical infrastructure and policy mechanisms for steering AI ecosystems toward robust, accountable, and democratic outcomes.

Speaker Biography: Benjamin Laufer is a PhD candidate at Cornell Tech, advised by Jon Kleinberg and Helen Nissenbaum. A recipient of a LinkedIn PhD Fellowship and three Rising Stars awards, Laufer researches how data-driven and AI technologies behave and interact with society. He previously worked as a research intern at Microsoft Research and a data scientist at Lime, and holds a BSE in operations research and financial engineering from Princeton University.

View the recording »

Computer Science Seminar Series

January 20, 2026

Abstract: Despite great potential, there is a growing gap between what AI systems promise and what they deliver, with real human costs. AI auditing is the practice of independently evaluating deployed AI systems to determine how they behave, what risks they pose, and whether they meet their intended objectives. This interdisciplinary endeavor requires both a technical expansion of our current AI evaluation paradigm and a framework for ensuring that audit investigations are sufficiently material for downstream legal actions and normative debates. At the intersection of law and public policy, applied economics and computer science, we can advance AI auditing policy and practice in material ways—by anchoring notions of engineering responsibility in AI development, expanding our vocabulary of AI evaluation methods, and pushing to connect AI audit outcomes to organizational and legal consequences. Through case studies of AI use in health care and government, we demonstrate how novel evaluation methods such as incident reporting, workflow simulations, and pilot experiments can supplement standard practices like data benchmarking to more adequately inform AI governance, shaping a range of outcomes from documentation and procurement to regulatory enforcement and product safety compliance. As auditing makes its way into key policy proposals as a primary mechanism for AI accountability, we must think critically about the necessary technical and institutional infrastructure required for this form of oversight to successfully enable safe widespread AI adoption.

Speaker Biography: Inioluwa Deborah Raji is a researcher at the University of California, Berkeley who is interested in algorithmic auditing. She has worked closely with industry, civil society, and within academia to push forward various projects to operationalize ethical considerations in machine learning practice and push forward benchmarking and model evaluation norms in the field. In particular, Raji aims to study how model engineering choices (from evaluation to data choices) impact consumer protection, product liability, procurement, anti-discrimination practice, and other forms of legal and institutional accountability related to functional harms. She is on the advisory boards for the Center for Democracy and Technology AI Governance Lab, the Health AI Partnership, TeachAI, REAL ML, and the Leadership Conference on Civil and Human Rights Center for Civil Rights and Technology. For her efforts, Raji has been named to Forbes’ 30 Under 30, MIT Technology Review’s Innovators Under 35, and TIME‘s 100 Most Influential People in AI lists. She is also the recipient of the 2024 Tech For Humanity Prize and the 2024 Mozilla Rise25 award, and is a co-recipient of the Electronic Frontier Foundation Pioneer Award along with Joy Buolamwini and Timnit Gebru. Raji received her bachelor’s of applied science in engineering science from the University of Toronto. She is currently completing her PhD in computer science at UC Berkeley.

View the recording »

Computer Science Seminar Series

January 20, 2026

Abstract: People with disabilities are marginalized by inaccessible social infrastructure and technology, facing various challenges in all aspects of their life. Conventional assistive technologies commonly provide generic solutions to a certain disability population and do not consider users’ individual and context differences, leading to high abandonment rate. Yuhang Zhao’s research seeks to thoroughly understand the experiences and needs of people with disabilities and create intelligent assistive technologies adaptive to user contexts, including their abilities, environments, and intents, providing effective, unobtrusive support tailored to user needs. In this talk, Zhao will discuss how she leverages state-of-the-art artificial intelligence, augmented reality, and eye-tracking technologies to design and develop context-aware assistive technologies. She will divide user context into external factors (e.g., surrounding environments) and internal factors (e.g., intents, abilities) and present her work on scene-aware, intent-aware, and ability-aware systems, respectively. Specifically, she will discuss: (1) CookAR, a wearable scene-aware AR system that distinguishes and augments the affordance of kitchen tools (e.g., knife blade vs. knife handle) for low-vision users to facilitate safe and efficient interactions; (2) GazePrompt, an eye-tracking-based, intent-aware system that supports low-vision users in reading; and (3) FocusView, a customizable video interface that allows users with ADHD to tailor video presentations to their sensory abilities. Zhao will conclude her talk by highlighting future research directions toward AI-powered context-aware systems for people with disabilities.

Speaker Biography: Yuhang Zhao is an assistant professor in the Department of Computer Sciences at the University of Wisconsin–Madison. Her research interests lie in human-computer interaction (HCI), accessibility, augmented/virtual reality, and AI-powered systems. Zhao leads the madAbility Lab at UW–Madison to design and build intelligent interactive systems to enhance human abilities. She has frequently published at top-tier conferences and journals in the field of HCI and accessibility (e.g., the ACM Conference on Human Factors in Computing Systems, the ACM Symposium on User Interface Software and Technology, the International ACM Special Interest Group on Accessible Computing Conference on Computers and Accessibility) and has received several U.S. and international patents. Her research has been funded by various agencies, including the NSF, the National Institutes of Health, the National Institute of Standards and Technology, and corporate sponsors such as Meta and Apple. Her work has received multiple Best Paper honorable mention awards and recognitions for contribution to diversity and inclusion and has been covered by various media outlets (e.g., TNW, New Scientist). Beyond paper publications, she disseminates her research outcomes via open-source toolkits and guidelines for broader impact. Zhao received her PhD in information science from Cornell University and her BA and MS in computer science from Tsinghua University.

View the recording »

Computer Science Seminar Series

January 15, 2026

Abstract: In 2025, frontier AI developers started warning that their AI systems were beginning to cross risk thresholds for dangerous cyber, chemical, and biological capabilities. This is unfortunate given how closed-weight AI systems are persistently vulnerable to prompt-injection attacks and open-weight systems are persistently vulnerable to malicious fine-tuning. Reinforcement learning from human feedback and refusal training aren’t enough. This presentation will focus on adversarial attacks that target model internals and their uses for making frontier AI safeguards “run deep.” In particular, we will focus on what technical tools can help us make open-weight AI systems safer. Along the way, we will discuss what AI safety can learn from the design of lightbulbs and why you should keep a close eye on Arkansas Attorney General Tim Griffin in 2026.

Speaker Biography: Stephen “Cas” Casper is a final-year PhD student at the Massachusetts Institute of Technology in the Algorithmic Alignment Group, where is he advised by Dylan Hadfield-Menell. Casper leads a research stream for the MATS Program and mentors for ERA and GovAI. He is also a writer for the International AI Safety Report and the Singapore Consensus on Global AI Safety Research Priorities. Casper’s research focuses on AI safeguards and governance, with features in the Conference on Neural Information Processing Systems; the Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence; Nature; the ACM Conference on Fairness, Accountability, and Transparency; the Conference on Empirical Methods in Natural Language Processing; the Institute of Electrical and Electronics Engineers Conference on Secure and Trustworthy Machine Learning; Transactions on Machine Learning Research; and the annual conference of the International Association for Safe and Ethical AI—as well as in a number of workshops and over 20 press articles and newsletters. Learn more on his Google Scholar page or personal website.

View the recording »

Computer Science Seminar Series

January 15, 2026

Abstract: AI evaluations inform critical decisions, from the valuations of trillion-dollar companies to policies on regulating AI. Yet evaluation methods have failed to keep pace with deployment, creating an evaluation crisis where performance in the lab fails to predict real-world utility. In this talk, Sayash Kapoor will discuss the evaluation crisis in a high-stakes domain: AI-based science. Across dozens of fields, from medicine to political science, Kapoor finds that flawed evaluation practices have led to overoptimistic claims about AI’s accuracy, affecting hundreds of published papers. To address these evaluation failures, he presents a consensus-based checklist that identifies common pitfalls and consolidates best practices for researchers adopting AI, as well as a benchmark to foster the development of AI agents that can verify scientific reproducibility. AI evaluation failures affect many other applications; beyond science, Kapoor examines how AI agent benchmarks miss many failure modes and presents systems to identify these errors. He examines inference scaling, a recent technique to improve AI capabilities, and shows that claims of improvement fail to hold under realistic conditions. Finally, Kapoor discusses how better AI evaluation can inform policymaking, drawing on his work on evaluating the risks of open foundation models and his engagement with state and federal agencies. Why does the evaluation crisis persist? The AI community has poured enormous resources into building evaluations for models, but not into investigating how models impact the world. To address the crisis, we need to build a systematic science of AI evaluation to bridge the gap between benchmark performance and real-world impact.

Speaker Biography: Sayash Kapoor is a computer science PhD candidate and a Porter Ogden Jacobus Fellow at Princeton University, as well as a senior fellow at Mozilla. He is a co-author of AI Snake Oil, one of Nature’s ten best books of 2024. Kapoor’s newsletter is read by over 65,000 AI enthusiasts, researchers, policymakers, and journalists. His work has been published in leading scientific journals such as Science and Nature Human Behaviour, as well as conferences like the Conference on Neural Information Processing Systems and the International Conference on Machine Learning. Kapoor has written for mainstream outlets including The Wall Street Journal and Wired, and his work has been featured by The New York Times, The Atlantic, The Washington Post, Bloomberg News, and many more. He has been recognized with various awards, including a Best Paper Award at the ACM Conference on Fairness, Accountability, and Transparency; an Impact Recognition Award at the ACM Conference on Computer-Supported Cooperative Work and Social Computing; and inclusion in TIME’s inaugural list of the 100 Most Influential People in AI.

View the recording »

Computer Science Seminar Series

January 13, 2026

Abstract: Governing AI is a grand challenge for society. Rishi Bommasani’s research provides the foundations for the scientific field of AI policy: How do we understand the societal impact of AI and how do we use our understanding to produce evidence-based AI policy? Bommasani’s research introduces new paradigms for measuring frontier models, deployed systems, and AI companies. Alongside his research, he will cover his work in multiple jurisdictions to demonstrate how AI research can impact public policy.

Speaker Biography: Rishi Bommasani is a senior research scholar at the Stanford Institute for Human-Centered AI researching the societal and economic impact of AI. His research has received several recognitions at machine learning conferences and has been covered by The New York Times, Nature, Science, The Washington Post, and The Wall Street Journal. Bommasani’s research shapes public policy: He is the lead author of the California Report on Frontier AI Policy that led to the first U.S. laws on frontier AI; he is an independent expert chair of the European Union AI Act General-Purpose Code of Practice, which clarifies the first comprehensive worldwide laws on frontier AI; and he’s an author of the International Scientific Report on the Safety of Advanced AI. Bommasani recently completed his PhD in computer science at Stanford University, where he was advised by Percy Liang and Dan Jurafsky and was funded by Stanford’s Gerald J. Lieberman Fellowship and the NSF Graduate Research Fellowship.