Emma Brunskill is an Associate Professor of Computer Science at Stanford University, with a courtesy appointment in Education. She holds a PhD in Computer Science from MIT (2009). Her research focuses on reinforcement learning, educational technology, and healthcare applications, aiming to develop AI systems that support human learning and decision-making. Notable projects include AI tutoring systems, policy evaluation methods, and behavior change interventions using large language models. Her work bridges theory and practice, addressing challenges in off-policy evaluation, fairness-aware decision making, and scalable educational tools. Brunskill has contributed to foundational research in reinforcement learning algorithms and their applications in real-world scenarios such as healthcare, education, and human-AI collaboration. She also leads initiatives to improve equity and efficiency in educational technologies through data-driven approaches. Brunskill's research has been supported by grants such as the NSF RI: Small grant for data-efficient reinforcement learning. She actively explores the ethical implications of AI systems, particularly in healthcare and education settings. Her recent work emphasizes leveraging large language models (LLMs) for personalized feedback and simulated training environments, as seen in studies like GPTCoach and LLM-based counselor upskilling.
Linyi Li is an Assistant Professor in the School of Computing Science at Simon Fraser University (SFU), leading the Trustworthy Artificial Intelligence (TAI) Lab. His research focuses on certifiably trustworthy deep learning systems, combining machine learning and computer security. He holds a PhD from the University of Illinois Urbana-Champaign (UIUC) and a B.Eng. from Tsinghua University. Affiliations: Simon Fraser University, TAI Lab Education: PhD in Computer Science, UIUC, 2023 B.Eng (Cum Laude), Tsinghua University, 2018 His research interests include deep learning , trustworthy machine learning , large language models , and software engineering . He emphasizes rigorous certification of robustness, fairness, and numerical reliability in AI systems. Recent work includes the InfiBench benchmark for evaluating code LLMs and advancements in neural network verification. Recent Research Trends: His publications span certified robustness, fairness guarantees, and scalable verification techniques for deep learning models. He also explores scientific evaluation of foundation models and adversarial defense mechanisms. Awards: Rising Stars in Data Science AdvML Rising Star Award Wing Kai Cheng Fellowship Finalist: Qualcomm Innovation Fellowship (2022) Winner: VNN-COMP'23 Competition (Team α, β-CROWN) Advising & Grants: As a PI, he oversees the TAI Lab's research. Though no specific grants are listed, his work is funded through competitive awards and university resources. Labs/Teams: Leads the TAI Lab at SFU, focusing on foundational and applied research in trustworthy AI.
Pan Xu is a tenure-track assistant professor with joint appointments in the Department of Biostatistics & Bioinformatics, Department of Computer Science, and Department of Electrical & Computer Engineering at Duke University. Previously, Xu was a Postdoctoral Scholar Research Associate at Caltech's Department of Computing and Mathematical Science and earned a Ph.D. in Computer Science from UCLA. Xu's research focuses on developing computationally- and data-efficient machine learning algorithms with strong empirical performance and theoretical guarantees. Xu's research interests center around Machine Learning with broad applications in Artificial Intelligence, Data Science, Optimization, Reinforcement Learning, and High Dimensional Statistics. The research specifically targets real-world problems in Bioinformatics and Healthcare, with recent work emphasizing distributionally robust decision making, efficient exploration strategies, and multi-agent systems. Xu has developed novel algorithms that address the challenges of exploration in sequential decision making and robustness to distributional shifts between training and deployment environments. Xu's recent publications demonstrate a strong trend toward developing theoretically grounded yet practical algorithms for reinforcement learning and bandit problems, with particular emphasis on distributionally robust methods, efficient exploration techniques, and applications to healthcare. The work spans both theoretical analysis (providing minimax optimal regret bounds) and practical implementations (validated on benchmarks like Atari games and real healthcare datasets). Whitehead Scholar award from Duke University School of Medicine (2023) Best Paper Award at ACM FAccT 2023 for Queer In AI paper PIMCO Postdoctoral Fellowship in Data Science (2022) TMLR Featured Certification (2023) NSF award on approximate sampling based exploration (2023) Xu actively mentors multiple Ph.D. students across Duke's Biostatistics & Bioinformatics, Computer Science, and Electrical & Computer Engineering programs, with several alumni now pursuing doctoral studies at top institutions. The research group has secured competitive funding including an NSF award for approximate sampling based exploration for sequential decision making. Xu serves as an action editor for TMLR and as an area chair for major conferences including ICML, NeurIPS, AAAI, ICLR, and AISTATS. Xu leads a dynamic research group focused on sequential decision making, with projects spanning theoretical algorithm development, implementation of practical systems, and applications to healthcare and bioinformatics. The group maintains active collaborations across Duke's medical and engineering schools, with recent work applying machine learning to epidemic forecasting during the pandemic.
Kwang-Sung Jun is an Assistant Professor at the University of Arizona, Department of Computer Science. His research spans interactive machine learning, reinforcement learning, and learning theory, with a focus on multi-armed bandits, Bayesian optimization, and generalized linear models. Education : Ph.D. in Computer Science from the University of Wisconsin-Madison (2015). Research Trends : Kwang-Sung's recent work (2023-2025) emphasizes bandit algorithms with second-order bounds, adaptive experimentation, and PAC-Bayes frameworks. He explores low-rank structures in regression, explainable reward shaping, and environmental risk modeling via probabilistic assessments of postfire debris-flows. His publications often bridge theoretical guarantees (e.g., regret bounds) with practical applications in machine learning and environmental hazards. Expertise : Interactive machine learning Multi-armed bandits Confidence sequences Reinforcement learning Human-machine hybrid systems
Sebastian Schemm is a Heisenberg Fellow at the Department of Applied Mathematics and Theoretical Physics (DAMTP), University of Cambridge, a position regarded as equivalent to a non-permanent Associate Professor. He leads research within the Atmosphere-Ocean Dynamics group and previously held an ERC Starting Grant-funded Assistant Professorship (without tenure track) at ETH Zurich. Education and Career Path PhD (2013) and MSc (2010), ETH Zurich, Switzerland Postdoctoral researcher, University of Bergen, Norway (2014–2017) Postdoctoral researcher, Laboratoire de Météorologie Dynamique, ENS Paris (2017–2018) Assistant Professor (ERC Starting Grant), ETH Zurich (2020–2024) Heisenberg Fellow, DAMTP, University of Cambridge (2025–present) Research Focus Schemm’s work centres on atmospheric and climate dynamics, spanning turbulence to planetary scales. Core themes include the physics of extratropical cyclone life cycles, jet-stream and storm-track dynamics, Rossby waves and teleconnection patterns, high-resolution atmospheric modelling, and the integration of machine-learning techniques for parameter estimation, data assimilation, and kilometre-scale global simulations. He also contributes to large-scale initiatives such as ECMWF’s WeatherGenerator. Scientific Awards and Editorial Service DFG Heisenberg Fellowship (2025) ERC Starting Grant (2020–2024) European Meteorological Society Young Researcher Medal (2019) Co-Editor, Weather and Climate Dynamics (EGU) Co-Editor, Quarterly Journal of the Royal Meteorological Society PhD Supervision & Funding He currently supervises PhD students at both Cambridge and ETH Zurich, with funding streams including the Cambridge CREATES Doctoral Training Partnership and Swiss/EU grants. Ongoing students explore reinforcement-learning parameterisations, jet-stream–storm-track relationships, mid-latitude eddy energetics, machine-learning ensemble forecasting, and Bayesian parameter estimation in LES. Active Projects EU Horizon project WeatherGenerator (led by ECMWF) PASC HiRAD-Gen : High-Resolution Atmospheric Downscaling Using Generative Models
Jeff Schneider is a Research Professor at the Robotics Institute within the School of Computer Science at Carnegie Mellon University. His research focuses on active learning, data mining, reinforcement learning, optimization, and intelligent control , applied to industrial and commercial challenges. Current PhD Students : Anoushka Alavilli, Benjamin Freed, Tejus Gupta, Albert Xu, Brian Yang Current Masters Students : Wen-Tse Chen, Xintong Duan, Aman Mehra, Vedant Mundheda, Zhouchonghao Wu Past PhD Students : J. Andrew Bagnell, Viraj Mehta, Matthew Tesch Past Masters Students : Ravi Tej Akella, Swapnil Pande, Siddharth Venkatraman Schneider's research bridges machine learning and autonomous systems , particularly in reinforcement learning , multi-robot coordination , and self-driving car technology . His work emphasizes practical applications of learning algorithms in real-world scenarios. His recent publications highlight advancements in offline reinforcement learning , multi-agent policy coordination , and behavior planning for autonomous vehicles . These studies often integrate deep learning and probabilistic modeling to address complex control and decision-making problems. Labs: Auton Lab CMU Center for Autonomous Vehicle Research Consulting & Industry Impact: Schneider has consulted for organizations like Uber ATG, Psychogenics, and Schenley Park Research, applying machine learning to domains such as self-driving cars , nuclear fusion , marketing , and drug discovery .
Georg Martius is a Full Professor in the Department of Computer Science at the University of Tübingen's Faculty of Science and a Max Planck Research Group Leader at the MPI for Intelligent Systems. Since April 2023, he has been a core member of the DFG-funded Cluster of Excellence 'Machine Learning: New Perspectives for Science,' which received extended funding through 2032 for its mission to integrate machine learning into fundamental scientific discovery processes. His academic foundation includes a PhD from the University of Göttingen and Bernstein Center for Computational Neuroscience (2005), a Diploma in Computer Science from the University of Leipzig (2003), and a visiting research period at the University of Edinburgh's Division of Informatics. Postdoctoral positions followed at the Max Planck Institutes for Dynamics and Self-Organization (Göttingen, 2009), Mathematics in the Sciences (Leipzig, 2010), and IST Austria (2015). Professor Martius's research pioneers the intersection of reinforcement learning, robotics, and tactile sensing, with emphasis on developing autonomous systems capable of natural locomotion, dexterous manipulation, and physical-world understanding. His work bridges theoretical machine learning with practical hardware applications, particularly in creating differentiable simulators, superresolution tactile sensors, and biologically plausible learning frameworks for robotic control. Analysis of his 2024-2025 publications reveals dominant trends in offline reinforcement learning (especially goal-conditioned and diversity-maximization techniques), object-centric representation learning for video understanding, and tactile sensing innovations. A strong thread connects foundation models to world model construction, while his work on differentiable physics engines enables precise collision handling and contact dynamics for real-world robotic control. His leadership roles include directing the Distributed Intelligence research team at Tübingen and contributing to major collaborative initiatives like the Real Robot Challenge and Myochallenge 2022. The Cluster of Excellence appointment represents recognition of his contributions to transforming scientific methodology through machine learning, particularly in automating hypothesis generation and experimental design. Current projects focus on integrating large-scale machine learning with embodied intelligence, advancing tactile perception systems like the Minsight vision-based sensor, and developing neuroplasticity-inspired approaches for robust out-of-distribution detection. His work directly impacts fields requiring physical interaction intelligence, from autonomous navigation to medical robotics, with emphasis on sample-efficient learning from limited real-world data.
Minsu Kim is a CIFAR AI Safety Post-doc Fellow at KAIST and Mila, collaborating with Prof. Yoshua Bengio, Prof. Sungjin Ahn, and Prof. Sungsoo Ahn. His work bridges System 2 Deep Learning, Bayesian posterior inference, and combinatorial optimization. Ph.D., Industrial Engineering, KAIST (2025) M.S., Electrical Engineering, KAIST (2022) B.S., Mathematics and Computer Science (Dual Degree), KAIST (2020) Kim's research focuses on enabling AI systems to measure uncertainty, represent causality, and perform sequential reasoning for safety-guaranteed planning. His methodology integrates GFlowNets and diffusion models with off-policy amortized inference, targeting applications in scientific discovery , hardware design optimization , and large language model alignment . Recent work explores Bayesian posterior inference through GFlowNets, aiming to unify deep learning with probabilistic reasoning. His 15 most recent publications (2025–2024) reveal a trend of combining combinatorial optimization with generative models for tasks like molecular graph discovery, vehicle routing, and neural architecture search. He also investigates diffusion samplers for Bayesian inverse problems and symmetry-based neural methods (Sym-NCO) to enhance sample efficiency. Jang Yeong Sil Fellowship (2025) KAIST Presidential Best Ph.D. Thesis Award (2025) Qualcomm Innovation Fellowship (2023) DesignCon Best Paper Awards (2021–2022) Kim's collaborations span KAIST's Industrial Engineering department and Mila's AI research groups. He actively contributes to academic peer review for top conferences (NeurIPS, ICML, ICLR) and journals (IEEE TNNLS, TPAMI), emphasizing the intersection of AI safety , uncertainty quantification , and systemic reasoning .
Lerrel Pinto is an Assistant Professor of Computer Science at New York University's Courant Institute, where he leads the General-purpose Robotics and AI Lab (GRAIL). His research focuses on enabling robots to generalize and adapt in unstructured environments through advancements in robot learning, decision making, and multimodal sensing. Before joining NYU, he completed a postdoc at UC Berkeley, a PhD in Robotics at Carnegie Mellon University, and an undergraduate degree in Mechanical Engineering at IIT Guwahati. Key research areas include large-scale robot learning, representation learning for sensory data, reinforcement learning for adaptability, and open-source robotics hardware. Notable achievements include the Sloan Fellowship (2025), NSF CAREER Award (2024), and Best Paper Awards at multiple robotics conferences. Pinto's lab has developed influential systems such as the AnySkin tactile sensing framework and the OPEN TEACH teleoperation system. Education highlights include a PhD from CMU (2019) under Abhinav Gupta, a postdoctoral stint with Alexei Efros and Pieter Abbeel at Berkeley, and undergraduate studies at IIT Guwahati. He has authored over 65 publications in top conferences like ICRA, NeurIPS, and CVPR. Pinto teaches courses on robotics, reinforcement learning, and AI at NYU. His service contributions include roles on program committees for ICML, NeurIPS, and IROS, as well as organizing workshops on topics like Dexterous Manipulation and Vision-Language Models for Robotics. His team actively collaborates through the GRAIL lab, with current projects exploring tactile sensing, zero-shot policy deployment, and multimodal robot learning systems. Ongoing research emphasizes bridging the gap between human and robotic dexterity through novel reward structures and adaptive control frameworks.
Dr. Meng Fang is a researcher specializing in Artificial Intelligence with a focus on Reinforcement Learning, Large Language Models, and their applications in medical QA, game theory, and causal inference. Their work combines technical innovation with practical problem-solving in safety-critical and domain-specific contexts. Key research areas: Social bias in AI, data augmentation, embodied agents, and model-based reinforcement learning Teaching: Coordinated module COMP532 - Machine Learning and BioInspired Optimisation (2024-25). Recent publications address challenges in offline RL robustness, vision-based safe reinforcement learning, and strategic game generalization.
Kamal H. Khayat serves as the Jones Professor of Civil Engineering at Missouri University of Science and Technology and directs the Center for Infrastructure Engineering Studies (CIES), focusing on advancing concrete technology for sustainable infrastructure development. His research spans high-performance concrete (HPC), ultra-high-performance concrete (UHPC), self-consolidating concrete (SCC), and concrete rheology, with specialized expertise in 3D printing applications, fiber reinforcement systems, and shrinkage mitigation techniques. He investigates innovative materials like superabsorbent polymers and alternative binders to enhance durability and sustainability in concrete infrastructure. Analysis of his recent publications reveals dominant trends in digital fabrication of concrete, particularly 3D printing optimization and rheological modeling for structural build-up. His work increasingly integrates machine learning for material property prediction while emphasizing eco-friendly formulations using recycled aggregates and carbon-mineralization techniques. As Director of CIES, Khayat leads multidisciplinary research initiatives in infrastructure materials engineering, overseeing projects related to concrete rehabilitation, sustainable construction methods, and advanced material characterization techniques for civil infrastructure systems.
Jiawei Zhang is an Assistant Professor in the Department of Computer Sciences at the University of Wisconsin–Madison since May 2025. His research focuses on optimization algorithms for machine learning, adversarial training, reinforcement learning, and distributed systems. Ph.D. in Computer and Information Engineering, Chinese University of Hong Kong, Shenzhen (2021) B.Sc. in Mathematics (Hua Loo-Keng Talent Program), University of Science and Technology of China His work spans nonconvex optimization , robust machine learning , and data-driven decision-making , with applications to AI and sustainable energy systems. Recent publications at ICML 2025 address stochastic primal-dual methods and contextual optimization robustness, while earlier works explore bilevel optimization, reward learning, and distributed consensus algorithms. Scientific awards include: MIT Postdoctoral Fellowship For Engineering Excellence (2023) CUHK-Shenzhen Presidential Award for Outstanding Doctoral Students (2021) SRIBD PhD Fellowship (2020-2021) CUHK-Shenzhen Outstanding Teaching Assistant (2023) He supervises undergraduate researchers and seeks graduate students with strong mathematical or algorithmic backgrounds for 2025 admission, emphasizing optimization and AI-driven applications.
Francis Y. Yan is an Assistant Professor of Computer Science at the University of Illinois Urbana-Champaign (UIUC), holding an affiliate appointment in Electrical & Computer Engineering within the Grainger College of Engineering. He leads the Illinois Networked Systems and AI (NSAI) research group, focusing on building intelligent networked systems that are safe, robust, and performance-optimized through practical machine learning integration. Prior to joining UIUC in January 2025, he served as a Senior Researcher at Microsoft Research Redmond under Victor Bahl. His educational background includes: Ph.D. in Computer Science from Stanford University (2020), advised by Keith Winstein and Philip Levis B.S. in Computer Science (Yao Class) and B.A. in Economics from Tsinghua University (2015) Additional undergraduate studies at MIT Yan's research adopts a holistic approach to practical machine learning for networked systems, emphasizing judicious application rather than indiscriminate use. He builds real-world systems and research platforms to lay ML foundations, devises deployable algorithms using domain insights, and validates performance through extensive empirical evidence. His work consistently addresses operator concerns regarding ML deployment—focusing on safety, robustness, generalization, and efficiency—while strategically combining ML with classical networking and systems techniques. Analysis of his 15 most recent publications (2023-2025) reveals dominant themes in resource allocation for microservices (DeDe, Autothrottle), real-time video optimization (Mowgli, GRACE), and LLM-driven network algorithm design. His work bridges theoretical advances with industrial deployment, evidenced by platforms like Puffer (400,000+ users) and OpenNetLab that have become community standards for validating congestion control algorithms. His research has been recognized with top honors: USENIX NSDI Outstanding Paper Award (2024) for Autothrottle APNet Best Paper Award (2022) IRTF Applied Networking Research Prize (2021) USENIX NSDI Community Award (2020) USENIX ATC Best Paper Award (2018) for Pantheon Yan actively recruits master's and undergraduate researchers for his NSAI group, prioritizing self-motivated students for projects in networked systems and AI. His research is supported by industry collaborations (notably Microsoft) and manifests in deployable platforms like Puffer—which has enabled award-winning research at NSDI and SIGCOMM—and OpenNetLab for real-time communications. His work directly impacts production systems including Microsoft Teams and Bing. He founded and directs the Illinois Networked Systems and AI (NSAI) research group, which operates critical infrastructure including Puffer (a live TV service and research platform) and OpenNetLab. These platforms facilitate community-wide validation of novel algorithms, with Puffer alone supporting multiple best-paper awards at top conferences. Current workstreams span cloud resource management (Teal, Autothrottle, DeDe), low-latency video (Puffer, Tambur, Mowgli), and LLM-augmented systems (Nada, Designing Network Algorithms via LLMs).
Anqi Liu is an Assistant Professor in the Department of Computer Science at the Whiting School of Engineering, Johns Hopkins University. She maintains significant affiliations with the Johns Hopkins Mathematical Institute for Data Science (MINDS) and the Johns Hopkins Institute for Assured Autonomy (IAA), while also collaborating extensively with the Center for Language and Speech Processing (CLSP) and the Laboratory for Computational Sensing and Robotics (LCSR). Her research focuses on developing principled machine learning algorithms for building reliable, trustworthy, and human-compatible AI systems in real-world applications. Key research areas include: Distributionally robust learning under covariate shift Uncertainty quantification for AI safety and fairness Safe exploration in control systems Fair machine learning under distribution shift Active learning under label shift Dr. Liu's work addresses critical challenges in high-stakes AI applications where reliability, safety, and societal impact are paramount. Her methods ensure AI systems remain robust to changing data environments, provide accurate uncertainty estimates, and incorporate human preferences in interactions. Analysis of her recent publications reveals a strong trajectory in trustworthy AI research with significant contributions to distribution shift handling, uncertainty quantification techniques, and safe decision-making frameworks. Her work bridges theoretical foundations with practical applications across healthcare, robotics, and social media analysis. Amazon Research Award Dr. Liu actively mentors eight PhD students and teaches specialized courses on Machine Learning for Trustworthy AI and standard Machine Learning at Johns Hopkins University, preparing the next generation of researchers to address critical challenges in AI safety and reliability.
Csaba Szepesvari is a Professor in the Department of Computing Science at the University of Alberta and Canada CIFAR AI Chair at Amii. His research focuses on developing efficient learning algorithms for sequential decision making problems, with particular emphasis on reinforcement learning theory, online learning, and bandit algorithms. Research interests include: Foundations of reinforcement learning and online decision making Convergence properties of learning algorithms Bandit problems and exploration-exploitation tradeoffs Function approximation in machine learning His publications demonstrate consistent theoretical contributions to understanding algorithmic convergence, complexity, and efficiency in reinforcement learning. Recent work explores LLM uncertainty estimation, policy gradient methods, and offline-to-online learning transitions. Leadership roles include: Foundations team lead at DeepMind Organizer of RL Theory Virtual seminar series He mentors graduate students in theoretical machine learning through the University of Alberta and Amii research programs.
