Find Academic Supervisors

Yizhou ZhangView profile

Assistant Professor

• Programming Languages
• Compiler Design
• Formal Methods
• +3 more

Yizhou Zhang is an Assistant Professor in the Department of Computer Science at the University of Waterloo. He holds a PhD and MS from Cornell University (2019 and 2016) and a BS from Shanghai Jiao Tong University (2012). His research focuses on programming languages, including design, implementation, and theory, with emphasis on formal methods, compiler optimization, and probabilistic programming. Education: PhD, Cornell University, 2019 MS, Cornell University, 2016 BS, Shanghai Jiao Tong University, 2012 Research interests span programming language theory, compiler construction, and formal verification. His work explores topics like certified compilers, effect handlers, and probabilistic program analysis. Recent publications emphasize formal models for memoization, nested family polymorphism, and bidirectional control flow. His publications reflect contributions to probabilistic programming semantics, compiler optimization techniques, and type systems. No scientific awards are explicitly listed. Advising and grant details are not provided in the text. Zhang’s research often intersects with formal methods and practical compiler implementation challenges.

• Machine Learning
• Artificial Intelligence
• Neural Networks
• +3 more

Daolang Huang is a Doctoral Researcher and Student in the Department of Computer Science at the School of Science, affiliated with Professor Samuel Kaski's group. He holds a Bachelor's degree in Engineering and Technology from Jinan University (2020). His research focuses on advanced machine learning techniques, including Bayesian inference, robust statistical modeling, and simulation-based methods. Key areas include experimental design optimization, neural processes, and equivariance in deep learning. Recent work emphasizes decision-aware algorithms and cost-effective simulation frameworks. Huang has collaborated internationally, with publications in top venues like NeurIPS. Despite no listed awards, his work demonstrates significant contributions to probabilistic modeling and optimization. Education : Bachelor's degree in Engineering and Technology, Jinan University (2020) Research Interests : Bayesian methods and amortized inference Robust statistics under model misspecification Continuous control and neural process architectures Optimization algorithms with decision-theoretic foundations Recent Research Trends : His articles (2020–2025) emphasize Bayesian experimental design, preference-based optimization, and equivariant neural networks. Themes include balancing statistical rigor with computational efficiency, particularly in high-dimensional decision-making contexts. Labs/Teams : Active member of Samuel Kaski’s research group, focusing on interdisciplinary applications of machine learning.

Yashar HezavehView profile

Associate Professor

• Dark Matter
• Machine Learning and Deep Learning
• Galaxy Clusters
• +3 more

Yashar Hezaveh is an Associate Professor at the University of Montreal's Faculty of Arts and Sciences, Department of Physics. He holds the Canada Research Chair in Astrophysical Data Analysis and Machine Learning. His work focuses on using gravitational lensing and machine learning to map dark matter distributions in galaxy halos, advancing our understanding of dark matter's nature. He completed his PhD at McGill University in 2013, earning recognition for groundbreaking research on high-redshift dusty star-forming galaxies. Education: PhD in Physics (McGill University, 2013) Affiliations: Kavli Institute for Theoretical Physics, Flatiron Institute's Center for Computational Astrophysics Research interests include applying deep learning to analyze gravitational lensing data, Bayesian neural networks for dark matter mapping, and cosmological simulations. Notable projects include the CASTOR mission and advances in radio interferometry image reconstruction. His work bridges astrophysics and machine learning, addressing challenges in cosmic structure analysis. Awards: Hubble Fellowship (2015), Top 10 Quebec Science Discoveries (2013). Grants: Leads multiple projects on dark matter, AI-driven stellar mass measurement, and astrophysical data analysis funded by NSERC, FQRNT, and the Simons Foundation. Students: Supervised four Master's theses on topics like Bayesian lensing inversion and machine learning for galactic archaeology. He contributes to collaborative initiatives like the Centre de recherche en astrophysique du Québec (CRAQ), fostering interdisciplinary astrophysics research.

Laurence Perreault-LevasseurView profile

Associate Professor

• Cosmology
• Machine Learning
• Gravitational Lensing
• +5 more

Laurence Perreault-Levasseur is an Associate Professor at Université de Montréal and an Associate Member of Mila. She specializes in applying machine learning methods to cosmology, with affiliations at the Flatiron Institute and Perimeter Institute. Her research focuses on gravitational lensing, dark matter, and precision cosmology. She holds a Canada Research Chair in Computational Cosmology and Artificial Intelligence. Education: PhD (University of Cambridge, 2015), M.Sc. and B.Sc. (McGill University). Research Interests: Machine learning for cosmological inference, strong gravitational lensing, galaxy cluster characterization, and dark matter studies. Affiliations: CRAQ (Québec Astrophysics Research Centre), Mila (Quebec AI Institute). Her work includes developing Bayesian methods for inverse problems and neural networks for astrophysical data analysis. Key projects involve precision cosmology via machine learning and reconstructing early-universe conditions using generative models.

• Reinforcement Learning
• Robotics
• Deep Learning
• +3 more

Glen Berseth is an Associate Professor in the Department of Computer Science and Operations Research at the University of Montreal and a Senior Academic Fellow at Mila – Quebec Institute for Artificial Intelligence. He is also a Canada CIFAR Chair in AI and Co-Director of the Montreal Robotics and Integrative AI Laboratory (REAL). His work focuses on reinforcement learning, robotics, and deep learning applied to autonomous systems. He holds a postdoctoral background from Berkeley Artificial Intelligence Research (BAIR), working under Sergey Levine. His research emphasizes real-world applications, including human-robot collaboration, continual learning, and multi-agent systems. He teaches courses on robot learning at the University of Montreal and Mila, covering cutting-edge techniques for general-purpose robots. Key research interests include reinforcement learning for robotics, adaptive interfaces, and sim-to-real transfer. His recent work addresses challenges in autonomous learning systems, such as robust locomotion control and efficient exploration strategies. Notable awards include the Canada CIFAR AI Chair. He has supervised numerous students, including PhD candidates Ozgur Aslan and Siddarth Venkatraman, and Master’s students like Roger Creus-Castanyer and Léa Demeule, focusing on topics like reinforcement learning and robotic control. Berseth leads research projects funded by organizations like the CRSNG, FCI, and MITACS, addressing topics such as modular lifelong learning and generalization in robotics. His lab, REAL, explores embodied AI and robotics integration.

François PottierView profile

Researcher

• programming languages
• functional programming
• type systems
• +5 more

François Pottier is a senior researcher at Inria in Paris, France, where he leads the Cambium research team. He is affiliated with the Gallium research group and has maintained an active research profile spanning several decades in programming languages and formal methods. His research focuses on programming languages , particularly functional programming and OCaml, type systems design and implementation, program verification using separation logic, and compiler construction . His work bridges theoretical foundations with practical implementation, evidenced by numerous open-source software tools he has developed. Pottier's publication record shows consistent contributions to top programming languages conferences (POPL, ICFP, ESOP) with recent work emphasizing separation logic frameworks, resource analysis, and formal verification of OCaml systems. His research demonstrates a clear trajectory from foundational type theory toward practical verification techniques for real-world programming languages. He has advised numerous PhD students including Remy Seassau, Tiago Soares, Clément Allain, and Alexandre Moine, among others, many of whom have gone on to research positions at institutions like New York University, Imperial College London, and Inria itself. Pottier maintains active service roles in the programming languages community as a member of IFIP Working Group 2.8 on Functional Programming and IFIP Working Group 2.16 on Language Design, and serves on program committees for major conferences including upcoming roles for JFLA 2026 and ITP 2026. He leads the Cambium research team at Inria, which focuses on programming language theory and implementation, particularly around OCaml and formal verification. The team develops both theoretical frameworks and practical tools that have influenced the broader programming languages ecosystem.

Ayush BhartiView profile

Research Fellow

• Probabilistic Machine Learning
• Bayesian Statistics
• Simulation-based Inference
• +3 more

Ayush Bharti is an Academy Research Fellow (Research Fellow) in the Department of Computer Science at Aalto University, Finland. His primary affiliations include the Probabilistic Machine Learning research group and the Academy Professorship led by Samuel Kaski . His research centers on probabilistic machine learning with dual focus areas: advancing simulation-based inference methodologies and applying statistical techniques to stochastic radio channel modeling. Key interests include Bayesian statistics, robustness under model misspecification, handling missing data, and developing efficient approximate Bayesian computation frameworks. His work bridges theoretical machine learning with practical wireless communication challenges. Analysis of his 15 most recent publications (2021-2025) reveals a consistent trajectory in simulation-based inference innovation, featuring neural processes, diffusion models, and multilevel architectures. Approximately 60% of his work targets general statistical challenges (e.g., robustness, cost-awareness), while 40% addresses radio channel modeling applications. Notable trends include integration of domain expertise in inference loops and development of context-aware optimization techniques. No information is available regarding student supervision or research grants. Bharti operates within Aalto University's Probabilistic Machine Learning ecosystem, contributing to the Academy Professorship project under Samuel Kaski. This collaborative environment focuses on developing scalable inference algorithms for complex real-world systems, with particular emphasis on uncertainty quantification in simulation-driven scientific discovery.

• Causal Inference
• Machine Learning
• Natural Language Processing
• +3 more

Dhanya Sridhar is an Assistant Professor at the Department of Computer Science and Operations Research (DIRO) of the University of Montreal and a core academic member of Mila - Quebec Artificial Intelligence Institute. She holds a Canada CIFAR AI Chair and co-leads the IVADO R3AI working group on safe and aligned AI. PhD from University of California Santa Cruz Postdoctoral research at Columbia University Data Science Institute Her research focuses on integrating causality and machine learning to build AI systems that are robust to distribution shifts, capable of efficient task adaptation, and aligned with human knowledge. Recent work includes causal effects of social interactions on US election participation and causal modeling in geriatric-oncology patient outcomes. Current research themes: Causal representation learning Temporal causal inference Counterfactual modeling Causal abstraction in large models Responsible AI development Scientific awards: Canada CIFAR AI Chair Major grants: NSERC Discovery Grant (PVX20965-RGP) 2023-2029 NSERC DGECR Grant 2023-2025 Multiple MITACS Acceleration Quebec grants Advising: PhD students: Philippe Brouillard, Shruti Joshi, Mizu Nishikawa-Toomey, Tom Marty, Cristian Manta

Stefan RadevView profile

Assistant Professor

• Bayesian Inference
• Deep Learning
• Probabilistic Modeling
• +3 more

Stefan Radev is an Assistant Professor in the Cognitive Science department at Rensselaer Polytechnic Institute . His work focuses on developing Bayesian methods with generative neural networks and computational models for complex systems like cognition and disease outbreaks. He is the core developer of the BayesFlow framework, which enables amortized Bayesian inference using deep learning. Radev's research addresses computational challenges in Bayesian workflows, such as rapid parameter estimation and model validation through neural networks. His primary research interests include Deep Learning , Probabilistic Modeling , and their applications in computational neuroscience and biomedical engineering. He collaborates with the Center for Modeling, Simulation and Imaging in Medicine (CEMSIM) and contributes to open-source projects like BayesFlow, which supports multi-backend frameworks (PyTorch/TensorFlow/JAX). Recent publications highlight advancements in amortized inference, simulation-based calibration, and robust Bayesian workflows. His work spans domains from cognitive modeling to biomedical imaging, emphasizing interdisciplinary applications of Bayesian methods. No scientific awards are explicitly listed, though his contributions to open-source tools and impactful research indicate significant academic recognition.

Justin DomkeView profile

Associate Professor

• Machine Learning
• Artificial Intelligence
• Probabilistic Models
• +3 more

Justin Domke is an Associate Professor in the Department of Computer Science at the University of Massachusetts Amherst, affiliated with the College of Computing and Information Sciences. He focuses on machine learning, probabilistic models, optimization, and statistical inference. His research addresses computational challenges in probabilistic reasoning and integrates approximate methods with complex models. He holds a PhD and MS in Computer Science from the University of Maryland, College Park, and dual BS degrees in Physics and Computer Science from Washington University in St. Louis. Domke has been recognized as an outstanding reviewer at CVPR 2011 and NIPS 2013. He is currently on leave for the 2023 Spring semester. His recent work spans variational inference advancements, including importance-weighted methods, Langevin diffusion approaches, and optimization techniques for hierarchical models. Domke has developed open-source tools like Pangolin and the Graphical Models Toolbox, and teaches courses on machine learning and graphical models at UMass.

• Computer Vision
• Machine Learning
• Statistics
• +3 more

Marcus A. Brubaker is an Associate Professor at York University's Department of Electrical Engineering and Computer Science within the Lassonde School of Engineering. He also holds an Adjunct Professor status at the University of Toronto's Department of Computer Science and serves as a Research Scientist at Google DeepMind. His primary affiliations include co-directorship of the Computational Vision and Imaging Lab (CVIL) at York University and membership in the Vector Institute as a Faculty Affiliate. Brubaker's research spans Computer Vision, Machine Learning, Statistics, and their applications to problems like cryo-electron microscopy (cryo-EM) and 3D molecular reconstruction. Education: He earned a Ph.D. in Computer Science from the University of Toronto (2011), with a thesis on physical models of human motion. His M.Sc. (2006) and B.Sc. (2004) in Computer Science were also from the University of Toronto. Research focuses include cryo-EM structure determination (notably the cryoSPARC software), normalizing flows, and probabilistic modeling. His work bridges computational imaging, statistical inference, and deep learning. Key contributions include advancements in cryo-EM algorithms, generative models, and camera noise analysis. He co-founded Structura Biotechnology and has contributed to open-source tools like Stan's math library. Publications emphasize novel view synthesis, neural radiance fields, sensor noise modeling, and cryo-EM reconstruction techniques. His research trends highlight integration of physics-based models with machine learning for solving complex imaging problems. Awards and recognition include impactful contributions to cryo-EM through cryoSPARC, a Nature Methods paper in 2017, and leadership in academic-industrial collaborations. Brubaker's labs and teams focus on advancing imaging technologies and their applications in structural biology and computer vision.

Wesley TanseyView profile

Assistant Professor

• Statistical Machine Learning
• Computational Oncology
• Bayesian Statistics
• +3 more

Wesley Tansey serves as Assistant Professor in the Computational Oncology group within the Department of Epidemiology and Biostatistics at Memorial Sloan Kettering Cancer Center (MSKCC). His research bridges statistical machine learning with cancer biology, focusing on developing novel computational frameworks for oncology applications. Dr. Tansey's research program centers on Bayesian statistical methods for biological data analysis, with particular emphasis on spatial transcriptomics (evidenced by his BayesTME framework), drug response modeling , and multi-omics integration . His lab develops scalable algorithms for high-dimensional biological data, including UnitedMet for metabolite imputation and MultiTME for spatial profiling analysis. Current projects address combinatorial drug screening optimization, tumor microenvironment characterization, and predictive oncology platforms for rare cancers. His recent publications (2023-2025) demonstrate strong focus areas: Bayesian active learning for drug screening (6+ publications) Spatial biology methods (BayesTME, MultiTME) Metabolomics-transcriptomics integration (UnitedMet) Causal inference in biological systems Scientific recognition includes serving as Area Chair for AISTATS 2022 and frequent invited talks at major conferences including SIAM's Mathematics of Data Science meeting. Dr. Tansey actively mentors lab members including Sophie Jaro (Spotlight presenter at ICML Workshop), Haoran Zhang (contributed talk presenter), Christopher Tosh (Associate Research Scientist), and Jeff Quinn (Bioinformatics Software Engineer). His lab receives research funding supporting development of computational oncology platforms with clinical translation potential. The VIVO Lab maintains active GitHub repositories for core methodologies including BayesTME, reflecting strong software engineering practices in computational biology.

• Probabilistic machine learning
• Bayesian experimental design
• Causal inference
• +4 more

Desi R. Ivanova is a research fellow at the University of Oxford's Department of Statistics under the Florence Nightingale Bicentennial Fellowship. Her work bridges probabilistic machine learning, Bayesian experimental design, and LLM evaluation frameworks. She holds a DPhil in Statistics from Oxford's StatML CDT program (2020-2024) and an MMORSE in Mathematics from University of Warwick (2011-2016) with Erasmus exchange at LMU Munich. Research spans causal machine learning and uncertainty quantification Developed CO-BED and Step-DAD frameworks Focus on LLM evaluation methodology and calibration Expert in real-time adaptive experimental systems Her publications demonstrate expertise in Bayesian self-consistency methods, neural data compression, and privacy-preserving dataset merging. Key contributions include improving amortized inference efficiency and developing gradient-based causal experimental designs. Current work emphasizes rigorous statistical evaluation of language models, advocating for appropriate uncertainty quantification when analyzing performance across small datasets. She critiques CLT-based methods for LLM evaluation and proposes more robust frequentist and Bayesian alternatives.

• Machine Learning
• Gaussian Processes
• Bayesian Inference
• +3 more

Paul Chang is a Visiting Professor in the Department of Computer Science at Aalto University, associated with the Professorship Solin A. He holds a Doctor of Technology (Tekn. toht.) degree in Computer Science from Aalto University, awarded in December 2024. His research focuses on machine learning, Gaussian processes, Bayesian inference, and sequential learning, with applications in neural networks and optimization. His work contributes to UN Sustainable Development Goals related to education and innovation through advancements in probabilistic modeling and computational methods. Chang's academic background includes a doctoral thesis titled Rethinking Inference in Gaussian Processes: A Dual Parameterization Approach (2024), exploring novel methods for improving efficiency and scalability in Gaussian process models. His research has been published in leading conferences such as ICML and ICLR, with notable contributions to sequential learning algorithms, Bayesian optimization, and memory-based approaches in Gaussian processes. Key research interests include dual parameterization techniques, sparse representations in neural networks, and the integration of simulation-based inference with diffusion models. His work bridges theoretical advancements in machine learning with practical applications in optimization, active learning, and temporal data analysis. Collaborations span international institutions, reflecting his global academic engagement. Chang's publications highlight advancements in function-space parameterization, memory-augmented Gaussian processes, and amortized probabilistic conditioning. These contributions address challenges in scalable Bayesian methods and sequential decision-making systems. His research emphasizes computational efficiency, robustness, and applicability to real-world problems in machine learning.

• Probabilistic Programming
• Deep Learning
• Artificial Intelligence
• +7 more

Jan-Willem van de Meent is an Associate Professor at the University of Amsterdam where he co-directs the AMLab with Max Welling. He also maintains an Assistant Professor position at Northeastern University, though currently on leave while continuing to advise students and collaborate. His research develops AI models by combining probabilistic programming and deep learning, focusing on understanding inductive biases that enable models to generalize from limited data. His research spans multiple domains: Probabilistic programming frameworks and inference methods Inductive biases for generalization from limited data Physical system simulators incorporating domain knowledge Causal structure and symmetries in AI models Applications in robotics, NLP, healthcare, and physical sciences Van de Meent is one of the creators of Anglican, a probabilistic programming language based on Clojure, and currently develops Probabilistic Torch, a library for deep generative models extending PyTorch. He is writing a book on probabilistic programming (draft available on arXiv) and serves as co-chair of the international conference on probabilistic programming (PROBPROG). His recent publications show a strong trend toward developing more efficient inference methods for probabilistic models, exploring disentangled representations across vision and language domains, and applying these techniques to healthcare, robotics, and neuroscience. His work on nested variational inference, energy-based models, and state abstraction in reinforcement learning has been particularly influential. Awards and Recognition NSF CAREER award (2021) Van de Meent actively advises multiple PhD students and postdocs across interdisciplinary projects. His lab maintains strong collaborations with researchers in robotics, healthcare, neuroscience, and other scientific domains, applying advanced probabilistic modeling to challenging real-world problems. He also develops practical tools for the research community, making advanced inference techniques more accessible to practitioners.