Pascal Poupart is a Professor and Canada CIFAR AI Chair at the Vector Institute, affiliated with the David R. Cheriton School of Computer Science at the University of Waterloo. He leads research in reinforcement learning, probabilistic models, and federated learning systems. Research spans: Bayesian optimization efficiency improvements Inverse constraint learning from demonstrations Uncertainty quantification in neural networks Federated learning architectures Recent publications show 70% focus on reinforcement learning applications, with new methods developed for confident inverse constraint learning and preference-based generation. Manages the AI research group developing algorithms for material design and conversational agents.
Andreas Vlachos is a Professor of Natural Language Processing and Machine Learning at the Department of Computer Science and Technology, University of Cambridge, and holds the Dinesh Dhamija Fellowship at Fitzwilliam College. His research spans dialogue modeling, automated fact-checking, imitation learning, semantic parsing, biomedical text mining, and trustworthiness in AI systems. PhD in Computer Science, University of Cambridge (supervised by Ted Briscoe and Zoubin Ghahramani) Lecturer at University of Sheffield Postdoctoral roles at UCL, University of Cambridge (NLIP group, Stephen Clark), and University of Wisconsin-Madison (Mark Craven) Current research focuses on evaluating and mitigating biases in language models, advancing fact-checking methodologies, and improving model robustness through interpolation, reinforcement learning, and causal reasoning. His work integrates natural logic, knowledge graphs, and multimodal evidence for verification tasks. Recent publications address uncertainty quantification, temporal planning benchmarks, and ethical framing of NLP artifacts. Grants from ERC, EPSRC, Facebook, Google, and the Alan Turing Institute fund his research team. Collaborations include Sebastian Riedel, Stephen Clark, and Mark Craven. Key projects explore disinformation detection, long-form generation, and confidence calibration in AI systems.
Zaid Harchaoui is an Adjunct Professor in the Department of Statistics at the University of Washington. His research focuses on machine learning, generative AI, and algorithmic optimization, with applications spanning ecology, neuroscience, and artificial intelligence. He explores learning under distributional shifts and develops tools for scalable generative models in language and vision domains. University: University of Washington Department: Statistics Research Focus: Learning from data with computational, inferential, and mathematical rigor; distributional shift adaptation; generative model scaling Email: zaid@uw.edu His recent work emphasizes generative AI applications in ecology and neuroscience, stochastic optimization for robustness, and algorithmic efficiency in large-scale learning. Key contributions include techniques for distributionally robust optimization, interpretable authorship obfuscation, and uncertainty quantification in behavior classification. Scientific awards and honors are not explicitly mentioned in the provided text. Collaborative efforts often intersect with nonlinear control algorithms, spectral analysis, and privacy-preserving machine learning frameworks.
Matthias Hein is a Professor at the Department of Computer Science, Faculty of Mathematics and Natural Sciences, University of Tübingen. His research focuses on Machine Learning , Adversarial Robustness , and Out-of-Distribution Detection , with applications in computer vision and medical imaging. He has received notable recognition including the Best Paper Honorable Mention Prize at ICLR 2021 and Outstanding Paper Award at CVPR 2021. His work includes developing benchmarks like RobustBench and Spurious ImageNet , and frameworks such as Sparse-RS and DIG-IN . His recent publications emphasize adversarial robustness across multiple domains (vision, text), counterfactual explanations for classifiers, and improved OOD detection methods . Collaborators include prominent researchers like Francesco Croce, Julian Bitterwolf, and Alexander Meinke. Scientific Awards : Best Paper Honorable Mention (ICLR 2021) CVPR 2021 Outstanding Paper Award Key Research Areas : Adversarial Robustness Vision-Language Models Medical Imaging AI Neural Network Calibration
Dr. Oana Cocarascu is a Senior Lecturer in Artificial Intelligence at the Department of Informatics, Faculty of Natural, Mathematical & Engineering Sciences, King's College London. She holds a PhD and MEng in Computing (Artificial Intelligence) from Imperial College London and conducts applied research focusing on how artificial intelligence can be deployed to support real-world applications, with machine learning and natural language processing as core components of her work. Her research interests span argument mining, explainable AI, machine learning, natural language processing, and symbolic reasoning. She is particularly focused on developing AI systems that can provide transparent, accountable, and ethical decision-making processes. Her work addresses critical challenges in bias mitigation, fairness metrics, fact verification systems, and argumentation-based explanations for complex AI decisions. Analysis of her recent publications (2023-2025) reveals a strong focus on fairness in AI systems, with particular attention to individual fairness metrics and nuanced evaluation frameworks. She has made significant contributions to fact verification systems, especially in multimodal contexts involving charts and tabular data. Her work increasingly integrates argumentation theory with natural language processing to create explainable AI systems that can justify their decisions through structured reasoning. Dr. Cocarascu leads an EPSRC-funded project titled 'A framework for evaluating and explaining the robustness of NLP models' (2024-2027) and is actively involved in the Natural Language Processing Group at KCL. Her research fingerprint shows strong activity in argumentation (100%), decision-making (74%), artificial intelligence (50%), explainable AI (39%), and bias mitigation (35%). She has contributed to numerous high-impact publications in top venues including ACL, EMNLP, AAAI, and the Journal of Artificial Intelligence Research, with a particular focus on making AI systems more transparent, accountable, and aligned with human values. Her work intersects with UN Sustainable Development Goals, particularly those related to reducing inequality and building resilient infrastructure.
Joel Goh is Associate Professor at the Department of Analytics and Operations, NUS Business School, National University of Singapore. He serves as Director of the J.Y. Pillay Comparative Asia Research Centre (under NUS Global Asia Institute) and PhD Program Director at the Institute of Operations Research and Analytics (IORA). Previously, he was Assistant Professor at Harvard Business School (2014-2017) and Visiting Scholar (2017-2022). BSc, MSc, PhD in Operations, Information, and Technology from Stanford University His research focuses on healthcare analytics (preventing health conditions, hospital operations, frailty assessment), supply chain analytics (digital business models, platform leakage), and service platform operations (hospital-at-home programs, incentive design). He co-created the Robust Optimization Made Easy (ROME) software. Recent publications analyze workplace psychological safety (2024), hospital-at-home models (2024), and platform leakage dynamics (2023). His work spans 18+ journals with 740+ citations for burnout cost studies (2022) and 606+ citations for physician well-being research (2017). Teaching Honors : 2023: Best MBA Teaching & Skinner Innovation Award 2021: NUS Annual Teaching Excellence Award 2020: Early Career Research Excellence Award & 40 Under 40 Best MBA Professors Advising & Grants : Served as PhD Program Director. Received NUS Start-Up Grant R-314-000-110-133 (2021) and Humanities & Social Sciences Fellowship (2021). Editorial roles include Associate Editor at Management Science , Manufacturing & Service Operations Management , and Senior Editor at Production and Operations Management .
Sendhil Mullainathan is the Roman Family University Professor of Computation and Behavioral Science at the University of Chicago Booth School of Business and a Professor of Economics and the Peter de Florez Professor of EECS at the Massachusetts Institute of Technology . His work bridges machine learning , behavioral science , and computational medicine , focusing on social problems like discrimination , poverty , and health equity . Research Interests : Behavioral economics, algorithmic fairness, poverty, AI in healthcare, and policy evaluation. Teaching : Courses on Artificial Intelligence and Algorithmic Solutions to Human Problems. Publications : Over 150 papers in journals like Science , Quarterly Journal of Economics , and Nature Medicine , with recent work on AI-driven healthcare disparities and behavioral economics. Scientific Awards : MacArthur ‘Genius’ Grant, Infosys Prize, ‘Top 100 Thinker’ (Foreign Policy Magazine), ‘Young Global Leader’ (World Economic Forum). Organizations : Co-founder of ideas42 (behavioral science non-profit), J-PAL (randomized trials in development), and Dandelion Health (healthcare data for AI). Serves on the MacArthur Foundation board.
Timothy M. Hospedales is a Professor of Artificial Intelligence at the Institute of Perception, Action and Behaviour within the School of Informatics at the University of Edinburgh . He also serves as VP AI and Head of Samsung AI Research Centre Europe . His research focuses on efficient and robust AI , emphasizing meta-learning , lifelong transfer-learning , and domain adaptation in both probabilistic and deep learning frameworks. Applications span computer vision , vision and language , reinforcement learning for robotics , and finance . Professor at University of Edinburgh (2020–present) ELLIS Fellow (2021) Head of Samsung AI Research Europe (2020–present) Founding Director of Applied Machine Learning Lab at QMUL (2012–2016) His work includes pioneering contributions to meta-learning , few-shot learning , and self-supervised methods , with notable awards such as the Best Paper Prize at ICML AutoML 2018 and Best Student Paper at ICPR 2018 . He has co-authored 15+ recent papers on topics like Vision-Language Models , Medical AI Fairness , and Diffusion Model Optimization . He served as Program Co-Chair for BMVC 2018 and AAAI 2022 , and authored a book on Visual Adaptation in the Deep Learning Era (2022). Co-Chair, BMVC 2018 Guest Editor, IET CV Special Issue (2016) Keynote Speaker at TASK-CV Workshop (ECCV 2016) Special Issue on Fewer Labels (IEEE PAMI 2020) His leadership extends to organizing workshops like the Learning-to-Learn Workshop at ICLR 2021 , Meta-Learning Workshop at NeurIPS 2020 , and Domain Generalisation Workshop at ICLR 2023 . Current projects include Meta-Omnium (CVPR 2023) for general-purpose meta-learning and MetaAudio (ICANN 2022) for few-shot audio classification benchmarks.
Karthik R. Narasimhan is a Professor at Princeton University's School of Engineering and Applied Science in the Department of Computer Science. Previously, he earned his PhD from MIT under Regina Barzilay and served as a visiting research scientist at OpenAI during 2017-18. His research focuses on the intersection of language and decision-making, building autonomous agents that learn from both experience and human knowledge. His research spans multiple high-impact areas including language agents (Text-DQN, CALM, ReAct, Tree of Thoughts), reinforcement learning (h-DQN, Multi-Objective RL), and AI safety (Toxicity in ChatGPT, DataMUX). He has developed critical datasets and benchmarks such as WebShop, InterCode, SWE-bench, and SILG that have become standard evaluation tools in the field. Current work emphasizes agent capabilities, software engineering automation, and multimodal interaction. His publication trends show strong focus on practical agent deployment (SWE-agent, Tree of Thoughts), safety evaluation (Probing AI Safety), and efficiency improvements (DataMUX). Recent work increasingly addresses real-world challenges in software engineering, security, and human-AI collaboration through rigorous benchmarking. Co-author of foundational GPT (2018) paper Key developer of Text-DQN (2015), CALM (2020), ReAct (2022), Tree of Thoughts (2023) Creator of influential benchmarks: WebShop (2022), SWE-bench (2023), InterCode (2023) He actively advises students through Princeton's computer science program, with research supported by multiple grants focused on autonomous agent development and language-based decision systems. His GitHub repositories (nlp-datasets, text-world-player) demonstrate strong community engagement in open-source research tools. Current projects include advancing language agent capabilities through Reflexion (2023) and Tree of Thoughts (2023) frameworks while addressing critical safety and efficiency challenges.
Dr Henry Moss is a Researcher at the Department of Applied Mathematics and Theoretical Physics within the School of Physical Sciences at the University of Cambridge. His work focuses on machine learning applications in climate modeling, Bayesian optimization, and Gaussian processes, bridging computational mathematics with environmental science and chemistry. His research interests include: Bayesian optimization for environmental and chemical systems Reinforcement learning in climate modeling Gaussian processes for molecular property prediction High-throughput machine learning in scientific domains Interpretable AI for coastal flooding prediction Hybrid ML-physics modeling Dr Moss's publications highlight his contributions to federated learning for climate models, sparse Gaussian process techniques, and multi-objective optimization frameworks. These works span applications in weather prediction, chemical engineering, and oceanography. Email: hwm26@cam.ac.uk
Tom Rainforth is an Associate Professor of Statistical Machine Learning at the University of Oxford's Department of Statistics, leading the RainML Research Lab (rainml.uk). He holds a Tutorial Fellowship at Mansfield College and is Principal Investigator of the ERC Starting Grant 'Data-Driven Algorithms for Data Acquisition' (2024–2029). Previously, he held roles including a postdoc under Yee Whye Teh (2017–2019), Junior Research Fellow at Christ Church College (2019–2019), and Florence Nightingale Bicentennial Fellow (2020–2024). He earned his MEng in Mechanical Engineering from the University of Cambridge and his D.Phil from Oxford under Frank Wood and Michael Osborne, focusing on probabilistic programming and Monte Carlo methods. He briefly worked in Ferrari's Formula 1 team. Research Interests : Bayesian experimental design, probabilistic and data-efficient machine learning, active learning, deep learning (with a focus on probabilistic approaches), probabilistic programming, and Monte Carlo methods. His work emphasizes statistical efficiency and adaptive algorithms. Publications : Recent contributions span modern Bayesian experimental design, adaptive importance sampling (Daisee), and applications of probabilistic methods in LLMs and generative models. His research bridges theory and practice, addressing challenges in scalability and robustness. Awards : ERC Starting Grant (2024–2029), highlighting his leadership in foundational AI research. Advising & Grants : Supervises 15 graduate students, including work on Bayesian neural networks, generative flows, and experimental design. His ERC grant supports cutting-edge data-driven algorithm development. Labs & Teams : Directs the RainML Lab, which develops scalable Bayesian methods and probabilistic AI systems.
David Duvenaud is an Associate Professor at the University of Toronto , holding a Canada Research Chair in Generative Models and a Schwartz Reisman Chair in Technology and Society . He is cross-appointed to the Department of Computer Science and Department of Statistical Sciences . A Sloan Research Fellow and founding member of the Vector Institute , his work bridges deep probabilistic models , AI safety , and scientific computing . PhD in Machine Learning (University of Cambridge, 2014) Postdoc in Hyperparameter Optimization (Harvard University, 2016) Co-founded Invenia (energy forecasting company) His research spans foundational Neural Ordinary Differential Equations (NeurIPS 2018 Best Paper) and Automatic Chemical Design (ACS Central Science 2018) to recent work on AGI governance (2025) and AI safety (2024). Key contributions include stochastic variational inference , implicit differentiation frameworks , and antisymmetrization layers for quantum Monte Carlo. Recent publications (2024-2025) focus on systemic existential risks from AI , many-shot jailbreaking attacks , and epistemic uncertainty quantification . His group trains energy-based models with scalable MCMC samplers and develops invertible neural architectures (e.g., Residual Flows NeurIPS 2019). He also explores human-AI alignment through LLM Processes (NeurIPS 2024) and Sycophancy in Language Models (ICLR 2024). Canada Research Chair (2025) NSERC Grant (2025) Sloan Research Fellow (2021) Schwartz Reisman Chair (2021) Best Paper Award (NeurIPS 2018) Distinguished Paper Award (ICFP 2021) His students include James Requeima , Jesse Bettencourt , and Raymond Douglas . He teaches courses on Statistical Methods for Machine Learning and Differentiable Inference . Current work (2025) investigates systemic human disempowerment through incremental AI capabilities and sabotage risk mitigation via hyperparameter-aware evaluations.
Mark Gales is Professor of Information Engineering at the University of Cambridge and an Official Fellow at Emmanuel College. He is currently on sabbatical leave for the 2024/25 academic year. Prior to his academic career, he worked as a consultant at Roke Manor Research Ltd, developing radar systems, before transitioning to speech and language processing. PhD in 'Model-Based Techniques for Robust Speech Recognition' (University of Cambridge, 1995) BA in Electrical and Information Sciences (University of Cambridge, 1988) His research focuses on speech and language processing , particularly in automated language assessment and low-resource speech technology . He leads the Automated Language Teaching and Assessment (ALTA) Institute , which collaborates with Cambridge University Press & Assessment (CUP&A) to develop commercial tools like Linguaskill and Speak & Improve . These platforms provide automated spoken/written assessment for millions of users globally. Recent publications highlight his work in LLM-driven speech processing , including adversarial attacks on foundation models, end-to-end spoken error correction, and uncertainty estimation frameworks. His team's research spans multilingual capabilities, with deployments in languages ranging from Dholuo to Tok Pisin . Awards : IEEE Fellow, ISCA Fellow Leadership : Fellows' Steward at Emmanuel College Mark has contributed extensively to Hidden Markov Model (HMM) applications in speech recognition, which underpinned early automatic speech systems. His work now bridges LLM-based language assessment with cross-lingual transfer learning and robustness testing for real-world deployments.
Baharan Mirzasoleiman is an Assistant Professor in the Department of Computer Science at the University of California, Los Angeles (UCLA), where she leads the BigML research group. Prior to joining UCLA, she was a postdoctoral research fellow in Computer Science at Stanford University working with Jure Leskovec. She received her Ph.D. in Computer Science from ETH Zurich advised by Andreas Krause. Her research focuses on addressing sustainability, reliability, and efficiency of machine learning, with particular emphasis on improving big data quality by developing theoretically rigorous methods to select the most beneficial data for efficient and robust learning. Her work spans several critical areas including data efficiency, robustness against label noise and data poisoning, and addressing spurious correlations in machine learning models. She has made significant contributions to understanding how neural networks exploit spurious features that correlate with certain categories during training but fail to generalize to minority groups. Professor Mirzasoleiman's research demonstrates how theoretically grounded approaches can lead to practical improvements in model robustness and efficiency across various applications including medical diagnosis and environmental sensing. Her work has resulted in the development of the SpuCo package, a Python library that provides modular implementations of state-of-the-art methods to address spurious correlations, along with controllable synthetic datasets like SpuCoMNIST and large-scale vision datasets like SpuCoAnimals. She has received numerous prestigious awards including the ETH medal for Outstanding Doctoral Thesis, being selected as a Rising Star in EECS by MIT, an NSF Career Award, a UCLA Hellman Fellows Award, and an Okawa Research Award. Her students have also received multiple fellowships and awards including Amazon Doctoral Student Fellowships and an OpenAI Superalignment Fast Grant. Professor Mirzasoleiman actively contributes to the academic community through invited talks at major conferences including ICML, ICLR, NeurIPS, and KDD, as well as co-organizing workshops on new frontiers in adversarial machine learning and sparsity in neural networks. She has developed educational resources including tutorials on Foundations of Data-efficient Learning presented at ICML 2024.
Joan Bruna is a Full Professor of Computer Science, Data Science, and affiliated Mathematics at New York University's Courant Institute and Center for Data Science. He leads the CILVR group and co-founded the MaD group. His research focuses on mathematical foundations of machine learning, deep learning, signal processing, and their applications in computational science, climate modeling, and geophysics. He holds a Ph.D. in Applied Mathematics from École Polytechnique and has held roles at UC Berkeley, the Institute for Advanced Study, and the Flatiron Institute. Education: Ph.D. (2013) and M.Sc. (2005) in Applied Mathematics, École Polytechnique and ENS Cachan; dual B.Sc./M.Sc. in Telecommunications and Mathematics from UPC Barcelona (2002–2004). Awards include the NSF CAREER Award (2019) and Alfred P. Sloan Fellowship (2018). Research interests span theoretical aspects of deep learning, optimization, and statistical methods, with applications to climate science and inverse problems. He has advised numerous PhD students and postdocs, contributing to seminal work on scattering networks, geometric deep learning, and neural ODEs. Professional service includes editorial roles at TMLR, JMLR, and TPAMI, plus program chairs for major conferences. His work bridges mathematics and machine learning, emphasizing rigorous analysis and real-world impact.
