David McAllester is a Professor at the Toyota Technological Institute at Chicago (TTIC) and holds a part-time Professor position at the University of Chicago's Department of Computer Science. He earned his B.S., M.S., and Ph.D. from MIT (1978, 1979, 1987). His research spans Artificial Intelligence, Machine Learning, and Theoretical Computer Science , with notable contributions to automated theorem proving (Ontic system), reinforcement learning, probabilistic programming, and computer vision. He is a Fellow of AAAI (since 1997) and has received multiple test-of-time awards for seminal papers in AI planning, constraint solving, and computer vision. Key Contributions: Developed the Ontic verification system for mathematical proofs. Pioneered conspiracy numbers in game tree search (influenced Deep Blue). Co-authored foundational work on policy gradient methods in reinforcement learning. Advanced PAC-Bayesian learning theory and co-training methods. Teaching: Teaches TTIC31230 (Fundamentals of Deep Learning), emphasizing mathematical rigor and research skills in computer vision, NLP, and reinforcement learning. Labs/Teams: Co-founded TTIC's research initiatives in AI and machine learning. Collaborates with industry and academia on foundational AI challenges. Awards: AAAI Fellow (1997) Test-of-Time Awards (AAAI, ICLP, CVPR)
Andrea Paudice is an Assistant Professor in the Department of Computer Science at Aarhus University. Their primary affiliation is with the Department of Computer Science, located at Åbogade 34, Building 5335, Room 317 in Aarhus N, Denmark. Paudice's research focuses on theoretical and algorithmic aspects of machine learning, optimization under uncertainty, adversarial robustness, and clustering methods. Research interests include developing robust statistical learning frameworks for heavy-tailed distributions, designing optimization algorithms with provable guarantees in stochastic settings, and exploring active learning strategies for efficient label usage. Recent work emphasizes high-probability bounds for stochastic methods, median-of-means techniques, and zeroth-order optimization under budget constraints. Publications span topics like adversarial noise mitigation, margin-based active learning, and exact cluster recovery via oracle queries. While no specific grants or awards are listed, their work demonstrates contributions to foundational machine learning theory and algorithmic robustness. No lab affiliations or student advising information is included in the provided text.
Maximilian Thiessen is a PhD student in machine learning at Technische Universität Wien , supervised by Thomas Gärtner. He is affiliated with the machine learning research unit and collaborates with the Laila lab in Milan. Research Interests : Learning with graphs Active learning frameworks Convexity theory in ML Computational learning theory Recent Research Trends include: (1) Expressive GNN architectures for outerplanar graphs (2025), (2) Generalized boosting theory through game frameworks (2024), (3) Efficient monophonic halfspace learning (2024), (4) Abstention mechanisms in contextual bandits (2024), (5) Global feature extensions in GNNs (2023), and (6) Expectation-complete graph representations (2023). Scientific Awards : 2024: DOC Fellowship from Austrian Academy of Sciences 2023: Best Poster Award at G-Research's ICML Poster Party Community Contributions : Organizer of Mining and Learning with Graphs (MLG) workshops at ECMLPKDD 2022-2024, co-organizer of Graph Learning on Wednesdays (GLOW) reading group, and session chair at ECMLPKDD'23.
Amit Mahajan, MD is an Associate Professor in the Department of Radiology & Biomedical Imaging at Yale School of Medicine. He specializes in neuroradiology with expertise in diagnosing head and neck cancers and performing procedures like spine biopsies. His academic roles include clinical practice and teaching at Yale University. Education & Training: MD from Armed Forces Medical College (1991) Residencies at All India Institute of Medical Sciences (1998) and Hospital of St. Raphael (2001) Fellowships in Radiology and Neuroradiology at Yale University School of Medicine (2003, 2010) Research Interests: Focuses on imaging techniques for brain tumors, immunotherapy response assessment, radiomics in oncology, and artificial intelligence applications in neuroradiology. Key areas include: Brain metastases and treatment efficacy PET/CT radiomics for cancer prognostication AI-driven tumor segmentation and radiomic feature extraction Publications Trends: Over 70 publications emphasizing translational research between clinical practice and bench science. Recent work explores combination immunotherapies, AI workflow integration, and imaging biomarkers for HPV-related cancers. Awards & Grants: No specific awards listed, but active in NIH-funded collaborative research projects. Serves on institutional review boards and editorial advisory roles. Labs & Teams: Collaborates with multidisciplinary teams in the Center for Brain & Mind Health and Stephen & Denise Adams Center for Parkinson’s Disease Research . Leads initiatives in imaging informatics and precision medicine.
Andres Masegosa is an Associate Professor at the Department of Computer Science, Aalborg University (AAU), within The Technical Faculty of IT and Design. He is actively involved in the DarkScience project (2022–present), focusing on metagenomic binning and microbial dark matter analysis. His research interests span Bayesian networks, machine learning, probabilistic graphical models, and educational methodologies in computer science instruction. Key research contributions include advancements in PAC-Bayes theory, genome representation learning, and cold posterior effects in Bayesian models. He has published extensively in top venues like Advances in Neural Information Processing Systems and Transactions on Machine Learning Research. His work often bridges theoretical contributions with practical applications in genomics and education. Masegosa leads the development of tools like InferPy for probabilistic modeling and has contributed to open-source projects such as the AMIDST toolbox. His educational research explores learning styles and active learning strategies, emphasizing live coding and programming exercises. Collaborations include interdisciplinary projects with microbiologists and data scientists, reflecting his expertise in computational methods for complex biological systems. His research portfolio demonstrates a strong focus on scalable probabilistic methods and their real-world applications.
Guillaume Ducoffe is an Associate Professor at the Faculty of Mathematics and Informatics, University of Bucharest, Romania, and a Senior Research Scientist at the National Institute of Research and Development in Informatics (I.C.I.), Romania. He is also affiliated with a joint research team between ICI and the Research Institute of the University of Bucharest (I.C.U.B.). Previously, he was a PhD student at Université Côte d'Azur, France, under the supervision of David Coudert, within the COATI project-team at Inria Sophia Antipolis. PhD, Université Côte d'Azur, France (2016) Master's Thesis, MPRI-ENS Cachan (2013) His research centers on algorithmic graph theory , with emphasis on computation in large graphs , including parameterized algorithms for problems like Diameter and Maximum Matching. He investigates metric tree-likeness in real-life networks through Gromov hyperbolicity, which has implications for routing efficiency and congestion. His work extends to information propagation using game-theoretic models such as coloring and hedonic games, and to online targeting detection , where he develops theoretically sound algorithms to uncover sensitive attribute targeting on the web via reductions to PAC learning of k-juntas. He also explores combinatorial topics like proper connectivity and Randic indices with applications in cryptography and chemistry. The analysis of his recent publications reveals a strong trend in theoretical computer science , particularly in the design and complexity of graph algorithms, structural graph properties, and their applications in network science and privacy. His work bridges pure graph theory with practical concerns in data centers, web transparency, and social networks. Guillaume Ducoffe has made significant contributions to both journal and conference literature, publishing in venues such as Discrete Applied Mathematics , SIAM Journal on Discrete Mathematics , ACM SIGMETRICS , and USENIX Security . His research is highly interdisciplinary, combining insights from algorithms, game theory, and network analysis. He actively supervises and mentors students, though specific names are not listed in the provided materials. He has been involved in research grants, including a postdoc grant from I.C.U.B., and has collaborated with prominent researchers such as David Coudert, Nicolas Nisse, Augustin Chaintreau, and Roxana Geambasu. His teaching includes core courses such as Data Structures and Algorithms , Advanced Graph Algorithms , and Advanced Programming Techniques at the University of Bucharest. Guillaume Ducoffe is a key member of collaborative research initiatives, including the joint ICI-I.C.U.B. team and the former COATI project at Inria. His work continues to advance the theoretical foundations of graph algorithms while addressing pressing issues in web transparency and network design.
Nathalie Huet is a Professor of Cognitive Psychology at the CLLE Laboratory (Cognition, Languages, Languages and Ergonomics), UMR 5263-CNRS, University of Toulouse Jean Jaurès. She serves as Director of the Department of Cognitive Psychology and Ergonomics within the Faculty of Psychology and holds leadership roles as Co-Scientific Manager of the CLLE Language and Cognitive Processes team and Co-Facilitator of the Education-Learning theme at CLLE. Her research centers on self-regulated learning models, examining relationships between cognitive, metacognitive, motivational variables and emotions in traditional and digital learning environments. She investigates virtual reality, augmented reality, and mixed reality applications (including Hololens 2) in academic and professional learning contexts, with emphasis on healthcare training for nurses and surgeons. Additional research strands include the use of AI aids like ChatGPT in learning processes and memory optimization factors related to language learning from an embodied cognition perspective. Analysis of her recent publications reveals a strong focus on embodied learning approaches in language education and memory research. Her work on embodied phonology methods for middle school English learners demonstrates significant impacts on pronunciation training and language acquisition. The PAC-PICL project represents a major contribution to embodied language teaching methodology, while her investigations into the "midscale disagreement problem" in psycholinguistic ratings have important methodological implications for cognitive science research. Professor Huet has supervised numerous doctoral students whose research aligns with her interests in self-regulated learning, embodied cognition, and educational technology. Her advisees have explored topics ranging from epistemic emotions in surgical training to cognitive investigations of pedagogical innovations and embodied approaches to language acquisition. She has secured various funding sources for these projects, including CIFRE partnerships with organizations like AFPA and SIMFORHEALTH. As Co-Scientific Manager of the Language and Cognitive Processes team at CLLE, Professor Huet leads interdisciplinary research that bridges cognitive psychology, linguistics, and educational technology. Her leadership in the CLLE Laboratory has fostered collaborations between psychologists, linguists, and ergonomists, creating a research environment focused on understanding human cognition in real-world educational contexts.
Oleg Stanislavovich P'yanykh is a Professor in the Department of Data Analysis and Artificial Intelligence at the Faculty of Computer Science, National Research University Higher School of Economics (HSE) in Moscow. He has been working at HSE since 2012 with 13 years of scientific and teaching experience. P'yanykh maintains active roles as a Guest Editor for both the American Journal of Roentgenology (since 2011) and Pattern Recognition (since 2010), and serves as a permanent member of the International DICOM Committee working group. His educational background includes a PhD in Computer Science from Louisiana State University (1998), an MS with honors in Applied Mathematics and Physics from Moscow State University (1994), and a Diplôme d'Etudes et de Recherches in Philosophy from French University College, Moscow State University and Sorbonne University (1994). P'yanykh's research spans medical informatics, image analysis and processing, medical information systems and standards (particularly DICOM and PACS), computer-aided diagnostics, teleradiology, machine learning, and control theory. His work bridges computer science with practical medical applications, focusing on improving medical imaging systems and data analysis. His publication record shows consistent contributions from 1997 through 2024, with recent work emphasizing scheduling algorithms, machine learning interpretability, human knowledge modeling, and medical image quality assessment. His research demonstrates a clear trajectory from foundational signal processing work to increasingly applied medical informatics solutions. Bonus for publication in an international peer-reviewed scientific journal (2019-2021) Bonus for publication in an international peer-reviewed scientific journal (2017-2018) Bonus for an article in a foreign peer-reviewed journal (2013-2015) P'yanykh has supervised student research including Viktor Sergeevich Lopatin's bachelor's thesis on 3D medical image processing and Ksenia Dmitrievna Loginova's master's thesis on medical image quality assessment. He has taught courses including Big Data and Machine Learning in Healthcare across multiple academic years (2020-2026) and Medical Informatics (2020-2022), reflecting his focus on applying computational methods to healthcare challenges.
Balder ten Cate is an Associate Professor at the University of Amsterdam, where he leads the Theoretical Computer Science research unit within the Institute for Logic, Language and Computation (ILLC). His academic journey includes previous positions at INRIA, UC Santa Cruz, Stanford, LogicBlox, and Google. He maintains active collaborations across institutions including the University of Bergen (co-advising PhD students) and Tsinghua University. Primary Affiliation: Theoretical Computer Science (TCS) Secondary Affiliation: Mathematical & Computational Logic (MCL) Office: Room L6.38, LAB42, Science Park 900, Amsterdam Dr. ten Cate's research spans diverse applications of logic in computer science and AI, with particular emphasis on data management, knowledge representation, and machine learning. His work bridges theoretical foundations with practical applications, focusing on how logical frameworks can enhance data systems and AI capabilities. He has developed significant contributions in finite model theory, database theory, and computational learning theory, with recent work exploring connections between universal algebra and learning theory. His publication record shows a strong trajectory in theoretical computer science, with recent papers focusing on query algorithms, interpolation in logical fragments, and the learnability of database queries. The research demonstrates consistent contributions at top venues including PODS, ICDT, IJCAI, and ACM Transactions journals. His work on 'Extremal Fitting Problems for Conjunctive Queries' received the ACM PODS 2023 Best Paper Award, while 'SAT-Based PAC Learning of Description Logic Concepts' earned an IJCAI 2023 Distinguished Paper Award. 2024 ETAPS Best Paper Nomination 2023 ACM SIGMOD Research Highlights Award 2023 Alberto Mendelzon Test-of-Time Award ICDT 2012 Best Paper Award 2006 EACSL Ackermann Prize for best dissertation Dr. ten Cate actively supervises PhD and MSc students, with recent theses covering topics like modal formula characterization, conjunctive queries, and temporal logic. He serves on numerous program committees including PODS 2025 and ICALP 2024, and has organized workshops on learning and logic. His current MSCA European Re-Integration Fellowship 'LLAMA: Logic and Learning: an Algebra and Finite Model Theory Approach' (2021-2025) demonstrates ongoing research leadership. He also contributes to academic service through committee memberships including the ASL Committee on Education and the editorial board of the Journal of Logic, Language and Information.
Liu Yang is a postdoctoral fellow in the Computer Science Department at Carnegie Mellon University , with a PhD from CMU under Avrim Blum and Jaime Carbonell. His research focuses on Theoretical Machine Learning and Theoretical Computer Science , exploring areas like Statistical Learning Theory , Property Testing , and Algorithmic Economics . He has contributed to active learning , transfer learning , and online pricing problems through mathematical frameworks. Research Interests : Liu's work bridges Computational Learning Theory with Algorithmic Economics , including: Mathematical theories for active property testing of Boolean functions Transfer learning with applications to online allocation and pricing Analysis of convex losses and statistical identifiability in learning Developing Buys-in-Bulk models for active learning efficiency Service and Teaching : He has served on program committees for ICML 2012-2013 , reviewed for top-tier venues, and taught courses like Graduate Algorithms and Modern Computer Algebra at CMU. He also co-developed the DistLearnKit MATLAB toolkit for distance metric learning.
Tim van Erven is an Associate Professor of Machine Learning at the Korteweg-de Vries Institute for Mathematics within the Faculty of Science at the University of Amsterdam. His research focuses on the mathematical foundations of machine learning, with particular expertise in online convex optimization, statistical learning theory, and explainable AI. He leads a research group dedicated to developing mathematically rigorous machine learning methods that work effectively without manual fine-tuning. His research interests span the mathematical foundations of machine learning, with emphasis on explainable machine learning, adaptive methods in online convex optimization, faster-than-minimax rates for 'easy data' in statistical learning, PAC-Bayesian concentration inequalities, and statistical learning theory with frequentist analysis of Bayesian methods. His work bridges theoretical guarantees with practical applications, recently shifting toward formal mathematical analysis of explainability methods for black-box AI systems. His recent publications reveal a clear evolution from foundational work in online learning and statistical theory toward explainable AI, with increasing focus on theoretical guarantees for concept learning and algorithmic recourse. The publications demonstrate strong methodological rigor while addressing practical challenges in interpretability and robustness of machine learning systems. Scientific Awards: VICI grant by the Dutch Research Council (2025) VIDI grant by the Dutch Research Council (2019) TOP grant by the Dutch Research Council (2016) NIPS 2014 outstanding reviewer award Rubicon grant by the Dutch Research Council (2011) Van Erven serves in significant academic leadership roles including as a member of the board of directors for COLT, co-chair for the AI & Mathematics initiative, and organizer of the thematic seminar on machine learning. He has successfully secured multiple competitive research grants and leads a research group working on the Mathematical Foundations for Explainable AI project, with several PhD and Postdoc positions currently open.
Nikita Zhivotovskiy is a tenure-track Assistant Professor in the Department of Statistics at the University of California, Berkeley. He previously held postdoctoral positions at ETH Zürich and Google Research, Zürich, and was affiliated with the Technion I.I.T. His academic background includes a PhD from the Moscow Institute of Physics and Technology, with affiliations during his studies at the Institute for Information Transmission Problems, Higher School of Economics, and Skoltech. His research lies at the intersection of mathematical statistics, probability theory, and learning theory . Key areas include robust estimation, online learning, statistical learning theory, algorithmic stability, and high-dimensional statistics. His work emphasizes theoretical foundations of machine learning, with a focus on generalization, risk bounds, and learning under non-standard assumptions. The recent publications reflect a strong trend in theoretical machine learning , particularly in understanding the limits and optimality of learning algorithms. Topics span PAC learning, online classification, private estimation, and clustering, often achieving dimension-free or high-probability guarantees. His work frequently appears in top venues such as NeurIPS, COLT, and FOCS, indicating significant impact in the field. Scientific Awards: Best Paper Award at Conference on Learning Theory (COLT), 2020 Nikita Zhivotovskiy has served as a reviewer for leading journals including Annals of Statistics , Probability Theory and Related Fields , and IEEE Transactions on Information Theory , and as a senior program committee member for COLT and ALT. He has co-taught courses at ETH Zürich and has advised or collaborated with numerous researchers, though formal students are not listed. His research has been supported through academic and industrial collaborations, including at Google Research. His work is embedded within the theoretical machine learning community, with active participation in workshops such as those at BIRS, and a growing body of work that bridges statistical theory and algorithmic design. While no formal lab is mentioned, his research group at UC Berkeley likely focuses on foundational aspects of learning and inference.
Daniel Liang is a Postdoctoral Researcher jointly supervised by Dr. Nai-Hui Chia at Rice University and Dr. Fang Song at Portland State University. His research bridges quantum computing, complexity theory, and learning theory, with a focus on efficient information extraction from quantum systems. PhD in Computer Science from University of Texas at Austin Bachelor of Science in Engineering from Cornell University in Computer Science and Engineering Physics His work explores time-efficient algorithms for quantum systems, aiming to benchmark quantum computers, discover new physics, and understand quantum computation limits. Key areas include stabilizer states, quantum tomography, and connections between learning theory and circuit complexity. Recent publications highlight advances in quantum state learning, stabilizer complexity, and pseudorandomness, with applications in tomography, optimization algorithms, and computational complexity reductions. ITCS 2023 Best Student Paper Award Collaborations span Rice University, Portland State University, and institutions like UT Austin, with interdisciplinary contributions to quantum algorithms and theoretical computer science.
Giuseppe Durisi is a Professor at Chalmers University of Technology in Gothenburg, Sweden, specializing in information theory and communication systems. His research bridges mathematically rigorous solutions with practical engineering applications in wireless and optical communication. Primary affiliation: Communication Systems Group , Chalmers University. Research focus: Optimal information transmission, 6G network design, and theoretical foundations of deep learning. Research Interests: Durisi investigates the interplay between latency, reliability, and throughput in digital communication, particularly in millimeter-wave and optical fiber channels . He develops finite-blocklength theory for efficient coding and explores how information theory can explain deep learning performance. Recent Article Trends: His 2025–2024 work emphasizes 6G distributed MIMO networks , energy-harvesting protocols , and machine learning integration into communication theory. Key themes include random access protocols , privacy in wireless aggregation , and hardware-constrained massive MIMO . Scientific Recognition: An IEEE Senior Member, Durisi has published extensively in top journals like IEEE Transactions on Communications and IEEE Transactions on Wireless Communications . Notable Collaborations: Work with teams on radio-over-fiber fronthaul , unsourced multiple access , and time-synchronized URLLC links .
Andreas Bjerregaard Jeppesen is a Research Fellow at the Department of Computer Science, University of Copenhagen, specializing in Machine Learning. His work bridges theoretical and applied research across diverse domains including quantum computing, biomedical informatics, environmental monitoring, and AI ethics. The Machine Learning section at DIKU explores foundational algorithms and their applications in Medical data analysis Remote sensing Biological modeling Sustainable AI Information retrieval . Andreas contributes to interdisciplinary projects like the SCIENCE AI Centre, leveraging the department's compute cluster for intensive simulations. His recent publications highlight trends in Quantum-inspired neural architectures Generative models for protein sequences Energy-aware AI development Neurological applications of ML Climate-conscious computing . Collaborations span computational biology, quantum chemistry, and federated learning for precision medicine.
