van der Schaar Lab

Inspiration Exchange

Inspiration Exchange is a series of engagement sessions aiming to share ideas and discuss topics that will define the future of machine learning in healthcare. These events will target machine learning researchers interested in healthcare, and will emphasise sharing of new ideas and development of new methods, approaches, and techniques.

As a lab, our purpose is to create new and powerful machine learning techniques and methods that can revolutionise healthcare. This doesn’t happen in a vacuum. At inception, we are inspired by ideas and discussions; in implementation, we need connections, trust, and partnership to make a real difference.

While you can learn about our work at major conferences in machine learning or in our papers, we think it’s a better idea to create a community and keep these conversations going. We’re also aware that many people—both in healthcare and machine learning—have questions about what we do, and how they can contribute.

For more information about Inspiration Exchange—and to sign up to join in—please have a look at the sections below, and keep checking for new updates.

Inspiration Exchange

Themed discussion sessions specifically for machine learning researchers in academia or industry interested in healthcare.

We would like to:
– discuss machine learning models and techniques
– share ideas about how machine learning can revolutionize healthcare
– spark new projects and collaborations
– raise awareness about this unique and exciting area of machine learning.

Standard session format:
– brief tutorial on chosen topic
– presentation of cutting-edge solutions by lab researchers
– roundtable discussions with industry leaders
– Q&A

The next session will be on September 16th 2026 at 4PM BST.

Join us for our next Inspiration Exchange on ‘From Learning Skills to Competing for Work: The Emerging Economy of AI Agents’. More details will follow soon.

The following sessions are: TBD

Time zones

Sign-up form

catch up on the most recent session (Session 56: June 26)

Session 1: introduction to automated machine learning

Ahead of this session, Mihaela published a piece of content on machine learning and the future of healthcare, entitled AutoML: powering the new human-machine learning ecosystem.

Session 2: automated machine learning pipelines

This session featured presentations from 4 of the van der Schaar Lab’s researchers regarding the Lab’s key AutoML projects. In addition, three session participants presented their own proposals for building on these projects by improving the AutoML methods or applying them in new contexts or applications.

Session 3: software packages for automated machine learning

This session featured presentations from 4 of the van der Schaar Lab’s researchers regarding the lab’s AutoML software packages (AutoPrognosis and Clairvoyance), followed by a Q&A.

Session 4: recent projects in machine learning for healthcare

This session covered a range of the lab’s most recent and exciting research projects in machine learning for healthcare. A total of 9 short presentations were given by the van der Schaar Lab’s researchers, followed by a Q&A session.

Session 5: synthetic data concepts and approaches

This session focused on synthetic data generation as one of the lab’s key research areas. Following an introduction from Mihaela van der Schaar, 3 short presentations were given by the van der Schaar Lab’s researchers, followed by a Q&A session.

Session 6: synthetic data evaluation

This session focused on evaluation of synthetic data. Following an introduction from Mihaela van der Schaar, 2 short presentations were given by the van der Schaar Lab’s researchers, followed by a Q&A session.

Session 7: application-oriented projects in machine learning for healthcare

This session featured 4 presentations (given by current and former lab members) on a range of application-oriented projects in machine learning for healthcare. Topics ranged from treatment effect estimation to multi-omics data integration, organ transplantation, and clinical trials.

A Q&A/open discussion took place in the latter half of the session, with participants asking researchers about their projects, and sharing their thoughts.

Session 8: individualized treatment effect inference (key concepts and challenges)

This was the first of 2 planned sessions focusing on individualized treatment effect inference (ITE inference).

The former half of the session featured 2 presentations introducing individualized treatment effect inference. This was followed by a participative discussion in which members of the lab and session participants discussed several key topics related to ITE inference, ranging from the evaluation of models to working with different types of data, to potentially relaxing some of the typical assumptions related to ITE inference.

Session 9: individualized treatment effect inference (time series)

This was the second in a pair of sessions focusing on individualized treatment effect inference (ITE inference). Building on our discussions the previous month, this session examined a range of ITE inference problems and approaches specifically within the time series setting.

The former half of the session featured 4 short presentations, followed by a participative discussion in which members of the lab and members of the audience discussed several key topics relevant to ITE inference in the time series setting, including counterfactuals, assumptions, missing data, and interpretability.

Session 10: quantitative epistemology

This session offered an introduction to quantitative epistemology—the van der Schaar Lab’s transformational new research area that aims to use AI and machine learning to understand and empower human learning and decision-making.

The former half of the session featured a high-level introduction to quantitative epistemology by Mihaela van der Schaar, followed by more in-depth explanations of specific frameworks and approaches by Ph.D. student Alihan Hüyük. The latter half of the session focused on a participative discussion in which Mihaela and Alihan answered questions from audience members about quantitative epistemology.

Session 11: time series in healthcare

The focus of the session was time series—an absolutely vital topic (and key research pillar for the van der Schaar Lab) that touches all aspects of machine learning for healthcare.

The first half of the session featured short presentations given by Mihaela van der Schaar, former Ph.D. student Changhee Lee, and research engineer Evgeny Saveliev. The latter half of the session featured a Q&A, in which Mihaela and members of the lab answered questions from audience members about time series in healthcare.

Session 12: what’s next for individualized treatment effects?

This session revisited the crucial topic of individualized treatment effect (ITE) inference, with a particular focus on looking to the future and introducing and discussing some brand-new approaches.

The first half of the session featured short presentations given by Ph.D. students Alicia Curth and Zhaozhi Qian. The latter half of the session featured a Q&A, in which Mihaela, Alicia, and Zhaozhi answered questions from audience members about their research and individualized treatment effects in general.

Session 13: how can we make ML models as robust and useful as possible?

This session raised a critically important question: how can we make ML models as robust and useful as possible?

Looking beyond predictions and analytics, there are a range of requirements and attributes that affect the degree of robustness and utility of a machine learning model—both in and beyond healthcare.

This session focused on three of these in particular: synthetic data, uncertainty estimation, and data imputation. Our lab members presented new and compelling work on each of these, while also conducting polls of the audience and asking attendees for their opinions and personal experiences.

Session 14: ML to transform organ transplantation

This was a particularly application-oriented session on ML tools for organ transplantation (one of the lab’s key research impact areas).

To start the session, clinician Alexander Gimson framed the importance and complexity of organ transplantation as a domain, and gave a quick overview of some of his collaborative projects with the lab to date.

Ph.D. student Jeroen Berrevoets then introduced OrganITE and OrganSync, two new ML-powered approaches to organ transplantation problems such as transplant benefit prediction, matchmaking, and queueing. Research engineer Bogdan Cebere presented a demonstrator for OrganBoard, a clinician-oriented survival prediction and matchmaking tool for organ transplantation. Finally, postdoc Fergus Imrie and PhD. student Yuchao Qin presented iTransplant, a machine learning project aiming to understand and improve decision-making around accepting and rejecting organs.

The session wrapped up with a short roundtable with a panel of clinical collaborators (all of whom are transplantation experts) and a brief audience Q&A.

Session 15: next frontiers in interpretability

This was an especially interactive session, featuring an audience poll and a lively Q&A and discussion.

To start the session, Mihaela van der Schaar shared a framework for understanding and categorizing the various different types of machine learning interpretability (in and beyond the healthcare setting).

Ph.D. students Jonathan Crabbé and Zhaozhi Qian then presented two fundamentally different ways of approaching interpretability (and related projects), and discussed their respective strengths. The session was then opened up to the audience for Q&A and general discussion.

Session 16: discovery using machine learning

This was an especially interactive session, featuring an audience poll and a lively Q&A and discussion. The session topic was “Discovery Using Machine Learning.”

To start the session, Mihaela shared her reasons for believing that machine learning can enable new discoveries (both in and beyond the healthcare setting) and why this is a new frontier for machine learning.

Ph.D. students Zhaozhi Qian and Krzysztof Kacprzyk then presented how machine learning can discover governing equations from data, highlighting a recent approach from our lab, and discussed the limitations of previous approaches. The session was then opened up to the audience for Q&A and general discussion.

Session 17: Causal Deep Learning

This was our first Inspiration Exchange session following a newly introduced format. This was an especially interactive session, featuring a roundtable discussion with industry guests, lively Q&A and discussions. The session topic was “Causal deep learning.”

To start the session, Mihaela shared a vision for combining causality and deep learning methods, which was further developed by Ph.D. student Zhaozhi Qian in an extended tutorial, where he presented the “Rung 1.5” framework for causality. Current and former lab members, Ioana Bica and Dr. Alexis Bellot, then presented two ways of adopting causality principles in generalisable imitation learning and causally aware imputation.

We then held our first roundtable discussion, where our industry guests Silvia Chiappa (DeepMind), Kim Branson (GSK), and Lindsay Edwards (Relation Therapeutics) shared their perspectives on practical challenges that can be addressed through causal deep learning. The session was then opened up to the audience for Q&A and general discussion.

Session 18: Causal Deep Learning Part 2

The session was a part 2 discussion on “Causal deep learning.”

To start the session, Ph.D. student Jeroen Berrevoets presented an extended tutorial, showcasing the “Rung 1.5” approach for imputation when estimating treatment effects. Subsequently, Ph.D. student Boris van Breugel introduced a method to adapt causality principles in generating fair synthetic data.

We then held our roundtable discussion, where our industry guests Amit Sharma (Microsoft Research), Cheng Zhang (Microsoft Research), and Jim Weatherall (AstraZeneca) shared their perspectives on practical challenges that can be addressed through causal deep learning.

The session was then opened up to the audience for Q&A and general discussion.

Session 19: ICML 2022 Preview

The session previewed the van der Schaar Lab’s accepted publications at the upcoming ICML 2022 conference and featured a roundtable discussion with industry guests, lively Q&A, and discussions. 

To start the session, there was a medley of presentations highlighting our 7 accepted papers, highlighting our lab’s continuing and impactful research. Our research encompasses a diverse array of topics, including discovering diverse classes of differential equations from data, machine learning interpretability, synthetic data, data-centric AI, AutoML, individualised treatment effects, and, last but not least, augmenting human skills using machine learning.

We then held our roundtable discussion, where our industry guests Dr Ari Ercole (University of Cambridge) and Dr Razvan Pascanu (DeepMind) shared their perspectives on ML and the medical impacts of the presented works.

The session was then opened up to the audience for Q&A and general discussion. 

Session 20: Neural differential equations

The session covered neural differential equations

To start the session, Ph.D. student Zhaozhi Qian introduced neural differential equations and motivated their use in the context of healthcare and medicine. Subsequently, Ph.D. student Samuel Holt presented Neural Laplace, a unified framework for learning diverse classes of differential equations by modeling dynamics in the Laplace domain. Additionally, Ph.D. student Nabeel Seedat presented Treatment Effect Neural Controlled Differential Equation (TE-CDE), which estimates counterfactual outcomes over time with irregularly sampled data. 

We then held our roundtable discussion, where our industry guests Dr Emma Slade (GSK) and  Dr Javier Hernández (Microsoft Research) shared their perspectives on ML and the medical impacts of the presented works. The session was then opened up to the audience for Q&A and general discussion.

The session was then opened up to the audience for Q&A and general discussion.

Session 21: Frontiers in ML Interpretability

Recording of the van der Schaar Lab’s 21st Inspiration Exchange session, which took place on October 25, 2022, and was attended by students and professionals from the AI and machine learning community.

To start the session, Prof. Mihaela van der Schaar presented a tutorial and introduction to ML interpretability, motivating their use in the context of healthcare and medicine. Subsequently, Ph.D. student Jonathan Crabbé presented his recent work on Concept Activation Regions, a generalised framework for concept-based explanations. Additionally, Ph.D. student Alicia Curth presented the ITErpretability benchmark, which utilises interpretability as a means to benchmark treatment effect models.

We then held our roundtable discussion, where our industry guests Dr Been Kim (Google Brain) and Dr Jianying Hu (IBM) shared their perspectives on ML and medical impacts of the presented works. The session was then opened up to the audience for Q&A and general discussion.

Sign up for our upcoming sessions: https://www.vanderschaar-lab.com/enga… The lab’s publications are here: https://www.vanderschaar-lab.com/publ…

Session 22: AutoPrognosis 2.0

In this Inspiration Exchange session, we introduced and discussed our lab’s latest tool – AutoPrognosis 2.0!

To start the session, Prof Mihaela van der Schaar gave an introductory presentation about AutoPrognosis 2.0, followed by a demonstration of the tool by Dr Fergus Imrie. Subsequently, Dr Thomas Callender (UCL) presented recent work utilising AutoPrognosis for lung cancer risk prediction. Core of the session was the roundtable discussion, where guests Dr Aditya Nori (Microsoft Health Futures), Prof Eoin McKinney (CCAIM), and Dr Thomas Callender (UCL) shared their perspectives on AutoPrognosis 2.0 and autoML in health. The session was then opened up to the audience for Q&A and general discussion. To learn more about AutoPrognosis 2.0, please see our recorded video, our dedicated website, and our paper.

Sign up for our upcoming sessions: https://www.vanderschaar-lab.com/enga… The lab’s publications are here: https://www.vanderschaar-lab.com/publ…

Session 23: Data-centric AI

This session was particularly engaging and interactive, with more than 120 participants joining the session live!

To start the session, Prof Mihaela van der Schaar gave an introductory presentation about Data-centric AI, followed by a presentation about our group’s work in the field by Dr Fergus Imrie, and a deep dive into Data-IQ by Nabeel Seedat. Core of the session was the roundtable discussion with our guest Prof Isabelle Guyon (Google Brain; INRIA) who shared her perspectives on Data-centric AI, challenges, and opportunities. The session was then opened up to the audience for Q&A and general discussion.

To learn more about Data-Centric AI, please see our recorded video, and see our dedicated website for Data-Check: Data-Centric AI Checklist.

Sign up for our upcoming sessions: https://www.vanderschaar-lab.com/enga… The lab’s publications are here: https://www.vanderschaar-lab.com/publ…

Session 24: Synthetic Data

This session was particularly engaging and interactive, with more than 140 participants joining the session live! 

To start the session, Prof Mihaela van der Schaar gave an introductory presentation about synthetic data. This was followed by a presentation by Dr. Zhaozhi Qian about our group’s latest open-source package Synthcity, which features a variety of advanced synthetic data generators.

The core of the session was the roundtable discussion with our panel of experts from healthcare, pharmaceuticals, consulting, and finance industries, including Dimitrios Vlitas (Accenture), Elena Sizikova (FDA), Nick Brown (AstraZeneca), Peter Krusche (Novartis), Robert Tillman (Optum Labs), and Tucker Balch (JP Morgan), who shared their perspectives on Synthcity and synthetic data. The session was then opened up to the audience for Q&A and general discussion.

To learn more about Synthcity, please see our recorded video below, our Github Repo, and our paper. We are also holding a hybrid synthetic data tutorial at AAAI-23 on Feb 8, 2023 at 2-6 PM EST, register here to learn more about Synthcity and synthetic data generation.

Session 25: Causal Deep Learning

In this session, Prof. Mihaela van der Schaar and Ph.D. student Jeroen Berrevoets unveiled our lab’s latest causal deep learning (CDL) manifesto (paper and slides attached). Please find below an abstract of our vision for the future of CDL:

Causality has the potential to truly transform the way we solve a large number of real-world problems. Yet, so far, its potential remains largely unlocked since most work so far requires strict assumptions which do not hold true in practice. To address this challenge and make progress in solving real-world problems, we propose a new way of thinking about causality – we call this causal deep learning. The framework which we propose for causal deep learning spans three dimensions: (1) a structural dimension, which allows incomplete causal knowledge rather than assuming either full or no causal knowledge; (2) a parametric dimension, which encompasses parametric forms which are typically ignored; and finally, (3) a temporal dimension, which explicitly allows for situations which capture exposure times or temporal structure. Together, these dimensions allow us to make progress on a variety of real-world problems by leveraging (sometimes incomplete) causal knowledge and/or combining diverse causal deep learning methods. This new framework also enables researchers to compare systematically across existing works as well as identify promising research areas which can lead to real-world impact.

The session also featured a lively roundtable discussion with our panel of experts Dr. Cheng Zhang (Microsoft), Dr. Guillermo Sapiro (Apple, Duke), and Dr. Simon Woodhead (Eedi), who shared their insights on the applications and developments of causal deep learning.

For more resources on causal deep learning, please see:

We encourage everyone to engage with this agenda, and together build the future of CDL!

Session 26: Revolutionizing Clinical Trials with Machine Learning

To start the session, PhD Student Alihan Hüyük gave an introductory presentation about clinical trials and how machine learning can help to decide when to make and break commitments to a trialing process. The core of the session was the roundtable discussion with our panel of experts, including Dr Carl-Fredrik Burman (AstraZeneca), Dr Miro Dudík (Microsoft Research), Dr Sean Hayes (Merck & Co.), and Prof Suetonia Palmer (University of Otago) sharing their insights on the applications and developments of machine learning for clinical trials. The session was then opened up to the audience for Q&A and general discussion.

For more resources on machine learning and clinical trials, see our research pillar on next-generation clinical trials and our perspective on how machine learning can revolutionize clinical trials.

Session 27: AISTATS & ICLR Preview 2023

As a recap, we had the following presentations:

Nabeel Seedat will present his work on how self-supervised learning can improve the quality of the conformal predictions for uncertainty quantification;

Alihan Hüyük will present Neural Laplace Control – a continuous-time model-based offline reinforcement learning method that combines a Neural Laplace dynamics model with a model predictive control planner;

Tennison Liu will present GOGGLE, a deep generative model for tabular data that exploits relational structure and an end-to-end message passing scheme to generate realistic synthetic tabular data;

Samuel Holt will present deep generative symbolic regression, which discovers closed-form mathematical equations from data by leveraging pre-trained deep generative models.

We encourage you to learn more about the presented work by reading the described papers. The titles, authors, abstract, and perceived impacts of all 11 accepted papers can be found on our lab website. 

Session 28: 2nd AISTATS & ICLR Preview 2023

Thanks so much to everyone who joined us at the recent Inspiration Exchange session, presenting our lab’s ICLR/AISTATS publications! You can now find the recording of the session on Youtube.

As a recap, we had the following presentations:

Alicia Curth presented her work on inferring heterogeneous treatment effects from time-to-event data in the presence of competing risks;

Boris van Breugel presented DOMIAS, a density-based membership inference attack model that aims to infer membership from synthetic data by detecting local overfitting of generative models;

Alan Jeffares presented TANGOS, a novel regularization method built on latent neuron attributions that encourages orthogonalization and specialization of neuron attributions to improve generalization performance;

Yuchao Qin presented T-Phenotype, a novel temporal clustering method that discovers phenotypes of predictive temporal patterns from time-series data.  

We encourage you to learn more about the presented work by reading the described papers. The titles, authors, abstract, and perceived impacts of all 11 accepted papers can be found on our lab website. 

Session 29: Treatment effect estimation and ML for Clinical Trials

As a recap, Alicia Curth presented three papers to be published at ICML 2023. Specifically, she discussed 1) how to conduct model selection in the absence of counterfactual information, 2) accounting for informative sampling when forecasting treatment effects over time, and 3) adaptively identifying populations with treatment benefits in clinical trials. 

For more, please see the publications:

On CATE model selection: https://arxiv.org/abs/2302.02923 

On informative sampling: https://arxiv.org/abs/2306.04255 

On adaptive trials: https://arxiv.org/abs/2208.05844

Additionally, the session was joined by Emre Kiciman (Microsoft Research) and Tom Diethe (AstraZeneca), who shared their insights on the developments in ML for treatment effect estimation and clinical trials.

Session 30: Data-Centric AI

Recording of the van der Schaar Lab’s 30th Inspiration Exchange session, which took place on 2 October 2023, and was attended by students and professionals from the AI and machine learning community.

In this jubilee session, we explore Data-Centric AI, a cornerstone in shaping robust and trustworthy machine learning systems. The quality of data used to train ML models determines the success or failure of AI. Data-centric AI provides a framework that brings data to the fore in designing ML pipelines. This is especially critical given the widespread use of data-hungry algorithms. Despite the importance of “data” nuances and particularities, there is little work that focuses on how to exploit them in our quest for designing robust and reliable ML. Hence, there is an urgent need to shift the focus to the data used in AI/ML and its quality.

We share recent advances and discuss new, unexplored directions that could yet provide new breakthroughs.

Additionally, we are excited to be joined by: Luis Oala (Head of Machine Learning, Dotphoton), Dr Manil Maskey (Senior Research Scientist, Office of Chief Science Data Officer, NASA), and Dr Peter Mattson (Senior Staff Engineer, Google; President of MLCommons.org), who will share their insights on new frontiers in data-centric AI.

Session 31: Synthetic Data and Simulators

In this session, we explored synthetic data and simulators. PhD student Boris van Breugel presented 3S-Testing, a deep generative modeling framework to facilitate model evaluation, by generating synthetic test sets for small subgroups and simulating distributional shifts. PhD Student Jeroen Berrevoets then discussed AllSim, a benchmarking environment for realistic simulation of the impact and utility of policies for resource allocation. Lastly, PhD student Nabeel Seedat shared his work on reimagining synthetic data generation through data-centric AI, exploring the potential of integrating data-centric AI techniques that profile the data to guide the synthetic data generation process.

Additionally, we were joined by Dr Maja Rudolph (Bosch Center for Artificial Intelligence), Dr Rianne van der Berg (Microsoft Research), and Dr Sergul Aydore (Amazon Web Services), who shared their insights on new frontiers of synthetic data and simulators. 

Session 32: Personalised Monitoring

In this session, we explore personalized monitoring. We discuss the importance of determining when and how we should monitor patients – an incredibly challenging yet vital aspect of medical care that requires fundamental advances in machine learning to address. Ph.D. student Samuel Holt presented “Active Observing in Continuous-time Control“, which addresses the problem of control in continuous-time environments while deciding when to take costly observations, a particularly relevant problem in real-world scenarios such as medicine. Then, Ph.D. student Yuchao Qin presented “Risk-Averse Active Sensing for Timely Outcome Prediction under Cost Pressure“, a novel risk-averse active sense approach that addresses the composite problem of when to conduct the acquisition and which measurements to make. 

Additionally, we were joined by Dr. Bill Weeks (Director, AI for Health Research, Microsoft) and Prof. Eoin McKinney (University of Cambridge), who shared their insights on new frontiers of personalized monitoring.

Session 33: Double Descent and Ensemble Learning

In this session, we explore double descent and ensemble learning, where we offer a fresh perspective on these advanced concepts, bridging the gap between theoretical understanding and practical applications in ML. PhD students Alicia Curth and Alan Jeffares presented their joint work “A U-turn on Double Descent: Rethinking Parameter Counting in Statistical Learning“ (NeurIPS23 Oral), examining claims of observed cases of double descent in classical statistical machine learning techniques, and whether they truly extend the limits of traditional U-shaped complex-generalisation curve. Then, Ph.D. student Alan Jeffares presented “Joint Training of Deep Ensembles Fails Due to Learner Collusion” (NeurIPS23 Poster), highlighting an intriguing effect of joint optimisation on the diversity and generalisation of deep ensembles.

Additionally, we were joined by Dr Jeremy Voisey (Senior Scientist, AI Research, Canon Medical Research Europe) and Dr Dimitrios Vitsios (Director, Data Science & Genomics, AstraZeneca), who shared their insights on the presented works.

Session 34: Large Language Models

In this session, we explored recent advances in LLMs, discussed innovative use cases of LLMs in medicine, and considered new, unexplored directions that could provide breakthroughs. Specifically, we discussed two works accepted for presentation at the upcoming ICLR 2024 conference. PhD students Samuel Holt and Max Ruiz Luyten presented their work “L2MAC: Large Language Model Automatic Computer for Unbounded Code Generation”, a novel approach using memory-augmented large language models (LLMs) for generating long and consistent outputs, here code, by overcoming the fixed context window limitation. Then, PhD students Tennison Liu and Nicolás Astorga presented “Large Language Models to Enhance Bayesian Optimization”, a novel framework that integrates the domain knowledge, in-context understanding, and learning capabilities of LLMs to improve the sample efficiency of Bayesian Optimization. 

Additionally, we were delighted to be joined by Prof Guillermo Sapiro (Health AI Lead, Apple & Duke) and Dr Hoifung Poon (General Manager, Microsoft Health Futures), who presented their thoughts and visions for LLMs.

Session 35: Innovative Use Cases of LLMs Part 2

In this session, we explored recent advances in LLMs, discussed innovative use cases of LLMs in medicine, and considered new, unexplored directions that could provide breakthroughs. Specifically, we discussed two recent works of the lab. PhD student Hao Sun presented their work “Query-Dependent Prompt Evaluation and Optimization with Offline Inverse RL” (ICLR 2024), a novel method that harnesses offline inverse reinforcement learning to draw insights from prompting demonstration data for prompt optimization. Then, PhD students Thomas Pouplin presented “Retrieval-Augmented Thought Process as Sequential Decision Making”, a novel process leveraging Monte-Carlo Tree Search to optimize retrieval to enhance thought processes.

Additionally, we were excited to be joined by Dr. Vivek Natarajan (Research Lead, Google Health AI)  and Dr. Chris Anagnostopoulos (Director of Life Sciences Innovation, QuantumBlack, AI by McKinsey), who shared their thoughts and visions for LLMs.

Session 36: Transforming Science with AI

In this session, we explore how AI can transform science. We share recent advances in machine learning methods for scientific discovery and demonstrate new ways of extracting meaningful and interpretable insights from real-world data. PhD student Krzysztof Kacprzyk presented their work “Shape Arithmetic Expressions: Advancing Scientific Discovery Beyond Closed-Form Equations”, a novel method that fuses GAM’s flexible shape functions with the complex feature interactions found in mathematical expressions. Then, Krzysztof presented “Towards Transparent Time Series Forecasting”, which proposes a top-down level of bi-level transparency for understanding time-series predictions.

Additionally, we were excited to be joined by Dr Ramon Hernandez (Global Head of Development Real World Evidence, Sanofi), Prof Richard Peck (Honorary Professor, Liverpool University; Associate Faculty, CCAIM), and Dr Sean Hayes (Associate Director, AI/ML, Merck & Co., Inc.) presenting their insights and perspectives on the potential for AI to transform science.

Session 37: Causality and Treatment Effect Estimation over Time

In this session, we present an overview of the state-of-the-art machine learning methods for estimating counterfactuals over time and explore open challenges for the field. We then motivate and explore a fundamentally new type of solution by reformulating the problem as an ODE discovery problem. Specifically, PhD students Krzysztof Kacprzyk, Samuel Holt, and Jeroen Berrevoets presented their work “ODE Discovery for Longitudinal Heterogeneous Treatment Effects Inference”. This approach has unique advantages over previous neural network-based architectures and could spark innovations in other areas of treatment effect estimation. 

Additionally, we are excited to be joined by Alexander D’Amour (Google DeepMind) and Dominik Janzing (Amazon Research), who share their insights on recent developments.

Session 38: Advanced topics in Diffusion Models

In this session, we provide an overview of the latest advancements in this field. We highlight some of our lab’s recent work presented at ICLR’24 and ICML’24, focusing on fundamental insights and improvements to diffusion models, including error propagation and expressiveness bottlenecks. Additionally, we explore time-series diffusion in the Fourier domain, uncovering impactful synergies between Fourier analysis and diffusion.

We are excited to be joined by Dr Max Ilse (Microsoft) and Dr Sander Dieleman (DeepMind), who share their perspectives on technical advances in diffusion models.

Session 39: Concept Bottleneck Models

In this session, we explore Concept Bottleneck Models, an innovative approach in machine learning that enhances model interpretability and performance by leveraging human-interpretable concepts. Van der Schaar Lab PhD Student Max Ruiz Luyten frames the topic by talking about our upcoming NeurIPS 2024 paper on the topic.

We are excited to be joined by Dr Hamid Palangi (Google) and Prof Eoin McKinney (University of Cambridge/CCAIM), who share their perspectives on Concept Bottleneck Models.

Session 40: Digital Twins

In this session, we explore Digital Twins, a novel area in machine learning focused on creating digital replicas of physical systems with advanced predictive capabilities. In particular, we will discuss how to automatically discover and build hybrid Digital Twins from data to model complex systems more effectively.

We are excited to be joined by Dr James Weatherall (AstraZeneca), Dr Jan Seidel (Boehringer-Ingelheim), Dr Jeremy Voisey (Canon Medical Research Europe), and Dr Thomas Klabunde (Sanofi).

Session 41: AI Model Evaluations and Deployments

In this session, we explore approaches to AI Safety and for ensuring reliable ML systems. We will explore novel ML approaches for model evaluation, as well as, deployment time adaptation of ML models. As part of the session, we present our upcoming NeurIPS 2024 papers “Context-Aware Testing: A New Paradigm for Model Testing with Large Language Models” and “Self-Healing Machine Learning: A Framework for Autonomous Adaptation in Real-World Environments“.

We are excited to be joined by Dr Jessica Schrouff (Google DeepMind) and Dr Grimaw Abebe Tadesse (Microsoft Research).

Session 42: LLM Co-Pilots

In this session, we explore the amazing opportunities LLM Co-Pilots offer in the empowerment of humans. We dive into exciting developments of LLM Co-Pilots and their synergy with researchers to solve task – in healthcare and beyond.

We are excited to be joined by Dr Bill Weeks (Microsoft), Dr Anders Boyd (Amsterdam UMC/Stitch HIV Monitoring), and Dr David Rivkin (UnitedHealth Group).

Session 43: Informed Meta-Learning and Active Task Disambiguation

In this session, we explore recent advances in LLMs. We discuss exciting developments of how LLMs can contribute towards Informed Meta-Learning and Active Task Disambiguation. To this end, we are presenting two of our papers accepted to ICLR 2025.

Our paper on Informed Meta Learning
Our paper on Active Task Disambiguation

We are excited to be joined by Dr Sebastian Farquhar (Google DeepMind), Dr Jane Wang (Google DeepMind), and Prof Andres Floto (University of Cambridge/CCAIM).

Session 44: Learning System Dynamics without Equations

In this session, we explore advances in Dynamical Systems with ML. Specifically, we dive into exciting developments in Learning System Dynamics without Equations.

We’re thrilled to welcome Dr Rich Caruana (Microsoft Research) as expert guest, a pioneer in transfer learning and specialist for medical decision making, transparent modelling, and deep learning.

To illustrate the state-of-the-art research on the topic, Krzysztof Kacprzyk present sour lab’s newest ICLR2025 paper, which builds on ideas and concepts we are introducing during the session!

Find our paper here.

Session 45: Genies – Next-Generation AI-Agents

In this session, we explore our newest vision: Genies – AI companions capable of dynamic reasoning, creative planning, and continuous learning.

Prof Mihaela van der Schaar explains what Genies are, what sets them apart from classic AI-Agents, and how they are able to empower humans in healthcare, business, and beyond.

We are excited to be joined by Prof Russ Altman (Stanford), and Dr Shakir Mohamed (Google DeepMind).

Learn more about Genies here.

Session 46: AI for Scientific Discovery

This session is comprised of two parts:

Part Ia: Overview of GenAI/LLM technologies for scientific discovery

Part Ib: Paper presentation – “Decision Tree Discovery Through LLMs via Semantically-Aware Evolution” (ICLR’25).

Part II: A discussion with our esteemed guests chaired by Prof. Mihaela van der Schaar.

For the discussion, we are excited to be joined by Dr. Joaquín Dopazo (Progress and Health Foundation & Institute of Biomedicine of Seville), and Dr. Markus Wulfmeier (Google DeepMind).

Session 47: AI Agents in Future Labor Markets

This session is comprised of two parts:

Part I: AI Agents in Future Labor Markets

Part II: A roundtable Q&A discussion with our esteemed guests chaired by Prof. Mihaela van der Schaar.

For the discussion, we are excited to be joined by Dr. Jim Weatherall (Chief Data Scientist, BioPharmaceuticals R&D, AstraZeneca), and Dr. Justine Rochon (SVP Global Biostatistics & Data Sciences, Boehringer Ingelheim).

Session 48: ICML Preview 2025

This session is comprised of two parts:

Part I: A preview of three papers, which will be presented at ICML 2025

– Strategic Planning: A Top-Down Approach to Option Generation presented by Max Ruiz Luyten.
– Preference Learning for AI Alignment: a Causal Perspective presented by Kasia Kobalczyk.
– G-Sim: Generative Simulations with Large Language Models and Gradient-Free Calibration presented by Samuel Holt.

Part II: A roundtable Q&A discussion with our esteemed guests chaired by Prof. Mihaela van der Schaar.

For the discussion, we are excited to be joined by Dr. Simon Woodhead (Chief Data Scientist, EEDI), and Dr. Rafik Salama (R&D Data and AI Transformation Lead, Accenture).

Session 49: Synthetic Data and Syncraft

This session will cover two distinct but complementary topics:

Syncraft, our no-code copilot for generating synthetic data – a powerful tool designed to accelerate experimentation and expand possibilities in data-driven research.

Frontiers and new developments in synthetic data – exploring the latest advances and opportunities in this fast-growing field.

We are excited to be joined by an outstanding panel bringing deep expertise • Alexandre Pasquiou – Co-Founder @Neuralk-AI • Juan Lavista Ferres – Corporate Vice President and Chief Data Scientist of the AI for Good Lab at Microsoft

Session 50: Autoformalism

This session celebrating the 50th session of IE covers Autoformalism. with talks from Tennison Liu and Prof Yuanzhang Xiao. Followed by a roundtable Q&A discussion with our esteemed guests chaired by Prof. Mihaela van der Schaar.

We are excited to be joined by an outstanding panel bringing deep expertise Prof Stefan Scholtes (Cambridge Judge Business School) Prof Radek Bukowski (University of Texas at Austin).

Session 51: NeurIPS Highlights 2025

In this session, we explore research from the van der Schaar Lab that will be presented at the upcoming NeurIPS 2025 conference (https://www.vanderschaar-lab.com/neurips2025/). 

We heard from the van der Schaar Lab members:

Claudio Fanconi: Cascaded Language Models for Cost-effective Human-AI Decision-Making
Qiyao Wei: Semantic-KG: Using Knowledge Graphs to Construct Benchmarks for Measuring Semantic Similarity
Silas Ruhrberg Estevez: Timely Clinical Diagnosis through Active Test Selection
Chris Chiu: Simulating Viva Voce Examinations to Evaluate Clinical Reasoning in Large Language Models

We are grateful for the following panellists to have joined the session:

Dr Lea Goetz – Senior AI/ML Engineer, Responsible AI, GSK
Dr Alison Smithard – Principal Scientist (AI Research team),Canon Medical Research
Dr Jeremy Voisey – Senior Scientist, Team Lead AI Research Bioinformatics, Canon Medical Research
Prof Eoin McKinney – Professor in the Department of Medicine, University of Cambridge

Session 52: AI in Biology

This session covered recently accepted AISTATS 2026 paper from the van der Schaar Lab, “Interpretable DNA Sequence Classification via Dynamic Feature Generation in Decision Trees.” You can read the paper here: https://openreview.net/pdf?id=433YQo03lM We were delighted to be joined by the paper’s authors who introduced the main findings and offered insights into how interpretable machine learning approaches can be applied to DNA sequence analysis.

  • Nicolas Huynh, PhD Student at van der Schaar Lab
  • Prof David Bentley FRS, Professor of Biochemistry and Molecular Genetics at the U. Colorado School of Medicine
  • Dr Ryan M Sheridan, Sr. Research Scientist at U. Colorado School of Medicine

Beyond the paper itself, the session discussed broader conversation around some exciting questions, including:

  1. As AI begins to automate large parts of data cleaning, pipeline construction, and exploratory analysis, what do you think will become the core contribution of bioinformaticians, and how should the field change to capitalize on AI/ML?
  2. How can we design ML systems where feature discovery is not only interpretable but also experimentally testable hypotheses that can directly drive wet-lab validation cycles?
  3. What would the ideal human-in-the-loop interface look like where bioinformaticians can guide, constrain, and interrogate AI-generated features during discovery rather than only after results are produced?

Session 53: Intelligence and Creativity Agenda in AI

This session explored a fundamental question for the future of artificial intelligence: are large language models truly creative, and if so, what kind of creativity do they exhibit?

While LLMs can generate novel ideas, text, and solutions, it remains unclear whether this reflects genuine creativity or the recombination of patterns learned during training. We will discuss different forms of creativity, how they might be defined or measured in AI, and what this might mean for the development of more advanced AI systems.

We were delighted to bring together two leading voices at the forefront of AI and creativity:

  • Vivek Natarajan – Research Scientist at Google DeepMind, leading research at the intersection of AI, science and medicine
  • Rafik Salama – Associate Director – UK Data & AI Lead for Life Sciences R&D at Accenture

Thank you to Max Ruiz Luyten – van der Schaar PhD student for his presentation on ‘The Reasoning–Creativity Trade-off: Toward Creativity-Driven Problem Solving’ https://arxiv.org/pdf/2601.00747 which was a recently accepted paper at AISTATS 2026.

Session 54: From Outputs to Processes: Re-thinking What Models Should Learn

This session explored a selection of accepted ICLR/AISTATS paper, from optimization paths, uncertainty formation, hierarchical abstraction, and planning, with each paper challenging a different “output-centric” assumption in ML. You can see our papers here: https://www.vanderschaar-lab.com/iclr-2026/ https://www.vanderschaar-lab.com/aistats-2026/

The following papers were featured:

Beyond the paper itself, we were thrilled to be joined by: Dr Jeremy Voisey – Acting Technical Manager, AI Research at Canon Medical Research Pierre-Yves Mousset – Principal Data Scientist at QuantumBlack, AI by McKinsey

Session 55: Scaling Causal Reasoning in LLMs

The session explored the challenges of scaling causal reasoning in machine learning systems. Causal reasoning requires answering interventional and counterfactual questions about the mechanisms that generate data, not just predicting what comes next. Yet today’s large-scale machine learning systems struggle in complex, real-world settings where causal structure is only partially specified and ground-truth supervision is limited. The talk highlighted that the core challenge is not simply one of model capacity, but of representation and supervision. We discussed a new perspective on scaling causal reasoning, one that reframes the problem as constructing and leveraging causal simulators capable of supporting interventional and counterfactual queries at scale, opening the door to new forms of supervision, evaluation, and generalization for causal reasoning in machine learning.

We are most grateful to our guests and speakers:

  • Dr Anna Gogleva – Senior Director Machine Learning at Relation
  • Dr Jonathan Richard Schwarz – Head of AI Research at Thomson Reuters, where he leads the Foundational Research Team, and is also a Visiting Professor at Imperial College London.
  • Anita Kriz and Nicolás Astorga – van der Schaar PhD students

The session is based on the recent paper “CauSim: Scaling Causal Reasoning with Increasingly Complex Causal Simulators” (https://arxiv.org/abs/2605.09079), led by Nicolás Astorga and Anita Kriz.

Session 56: AI for Bio-Science Discovery

The session explored emerging directions at the intersection of AI and biological discovery, building on one of the key research frontiers represented in the van der Schaar Lab ICML 2026 accepted papers. Will featured the following 3 papers:

We are most grateful to our guests and speakers:

  • Prof David Bentley FRS – Professor of Biochemistry and Molecular Genetics at the U. Colorado School of Medicine
  • Prof Gerard Evan FRS – Group Leader of Cancer Biology Laboratory, at The Francis Crick Institute
  • Dr Roderik Kortlever – Research Fellow in the ‘Cancer Biology Laboratory’ Group of Prof. Gerard Evan, at The Francis Crick Institute & King’s College London
  • Dr Rory Maizels – Postdoctoral Fellow at EMBL EBI & the Sanger Institute

To learn more about our team of researchers, click here. You can find our publications here.

Upcoming events

No event found!