van der Schaar Lab

Inspiration Exchange 2022

As the end of the year approaches, we wanted to recap an exciting year of engagement and take a moment to thank everyone who attended Inspiration Exchange in 2022. Our audience, their engagement, and contributions have made this series a truly enriching experience for all of us.

In addition, we are grateful to have had the opportunity to host such a diverse group of speakers, who shared their knowledge and insights with us on a wide range of topics.

You can find recordings of all our sessions here. There, you can also find the option to sign up for our future sessions.

Our 2022 sessions

ML to transform organ transplantation (7th Jan) – We began the year with an applications-focused session for ML in organ transplantation, discussing recent work OrganITEOrganSync, and iTransplant. We were joined by 3 clinicians working in organ transplantation Dr Alexander Gimson, Dr Vasilis Kosmoliaptsis (University of Cambridge), and Dr Brent Ershoff (UCLA).

Next frontiers in interpretability (7th Feb) – In this session we discussed the unique requirements for interpretability in healthcare, highlighting the multi-faceted nature of the problem and the different types of interpretability needed.

Discovery using machine learning (4th Mar) – Machine learning is capable of more than building predictive models – in this session, we explored how ML can be used to discover governing equations from data.

Causal deep learning (Part 1 – 14th AprPart 2 – 27th May) – These were our first Inspiration Exchange sessions with a newly introduced format! Across two sessions, we discussed the emergent field of causal deep learning and the impact principles of causality could have in machine learning. In Part 1, we were joined by Dr Silvia Chiappa (DeepMind), Dr Kim Branson (GSK), and Dr Lindsay Edwards (Relation Therapeutics), and in Part 2, we welcomed Dr Amit Sharma (Microsoft Research), Dr Cheng Zhang (Microsoft Research), and Dr Jim Weatherall (AstraZeneca) for lively “round-table” discussions.

ICML 2022 preview (21st Jun) –  We were delighted to be joined by Dr Razvan Pascanu (DeepMind) and Dr Ari Ercole (University of Cambridge) as we previewed our lab’s work at ICML.

Neural differential equations (4th Oct) – Neural differential equations are a fascinating new class of machine learning models. In this session, we provided an introduction to neural DEs and discussed their impact in healthcare and beyond with Dr Emma Slade (GSK) and Dr Javier Hernández (Microsoft Research).

Frontiers in ML interpretability (25th Oct) – We were thrilled to be joined by Dr Been Kim (Google Brain) and Dr Jianying Hu (IBM) as we revisited the critical topic of interpretability in healthcare, presenting recent advances in concept-based explanations and an example of how interpretability can be used to evaluate models.

AutoPrognosis 2.0 (22nd Nov) – In this session, we introduced AutoPrognosis 2.0, our lab’s latest open-source software package that allows users to build diagnostic and prognostic models using automated machine learning. We discussed the role and impact of such approaches in medicine with Dr Aditya Nori (Microsoft Health Futures) and clinical collaborators Prof. Eoin McKinney (University of Cambridge) and Dr Thomas Callender (UCL), the latter presenting an early application of AutoPrognosis 2.0 to lung cancer.

Data-centric AI (13th Dec) – Our final session of the year explored data-centric AI with Prof Isabelle Guyon (Google Brain), who recently gave a fascinating Keynote at NeurIPS 2022 on the data-centric era. In this session, we discussed a framework for data-centric ML as well as specific methods for understanding training and test data.

Our next session will be held on 1 February 2023. The focus of the session will be Synthetic Data, a new and exciting area of machine learning with real-world impact.

For more on our work on synthetic data, see our research pillar and ICML 2021 tutorial.

More details to follow in the New Year!

We are very much looking forward to welcoming you to our Inspiration Exchange in 2023!

Tennison Liu

Tennison is a Ph.D. student in the Cambridge Centre for AI in Medicine as well as the van der Schaar Lab.

Tennison graduated from the University of Sydney with a B.Eng in Electrical Engineering, receiving the University Medal (highest mark), and then continued to an M.Phil. in Machine Learning and Machine Intelligence at the University of Cambridge, where he first worked with Prof. van der Schaar and was awarded the John CB Chau Prize for highest M.Phil. mark.

Tennison has held research data scientist roles at Cochlear, IBM, and Macquarie Group in Australia, but felt drawn to the intellectual stimulation of machine learning research. In his own words, he opted to join the lab for its “great resources that are rarely found in other institutions, in terms of research collaborations, faculty, and brilliant students.”

Tennison currently expects his work with the van der Schaar lab and the Cambridge Centre for AI in Medicine to focus on areas such as synthetic data, discovery using machine learning, and deep self-supervision.

In his spare time, Tennison enjoys rowing with the St Edmund’s College Boat Club, playing basketball, and working out in the gym. He also enjoys photography and (in the right mood) will play the piano and trumpet.

Tennison’s research is supported by funding from AstraZeneca.

Andreas Bedorf

Fergus Imrie

Fergus Imrie is a postdoc at the ECE Department, University of California, Los Angeles (UCLA).

He is excited and motivated by the promise of transforming healthcare and improving medical knowledge through the use of machine learning in combination with clinical experts.

In particular, Fergus is interested in self-supervised learning and methods for understanding clinical decision making.

Prior to joining the lab, Fergus completed his DPhil (PhD) at the University of Oxford in the Department of Statistics, developing deep learning approaches for drug discovery. Fergus values work with a strong translational impact: his research is currently being used by a number of pharmaceutical companies on active drug discovery projects to develop new therapeutics.