Publications and Talks


*denotes equal contribution


Speedy Performance Estimation for Neural Architecture Search. Robin Ru*, Clare Lyle*, Lisa Schut, Miroslav Fil, Mark van der Wilk, Yarin Gal. NeurIPS 2021 Spotlight.

[ArXiv pre-print]

Generating Interpretable Counterfactual Explanations By Implicit Minimisation of Epistemic and Aleatoric Uncertainties. Lisa Schut*, Oscar Key*, Rory McGrath, Luca Costabello, Bogdan Sacaleanu, Medb Corcoran, Yarin Gal. AISTATS 2021.

[Paper] [Code]

A Bayesian Perspective on Training Speed and Model Selection. Clare Lyle, Lisa Schut, Robin Ru, Yarin Gal, Mark van der Wilk. NeurIPS 2020.

[Conference proceedings] [Arxiv Paper] [Clare's Blogpost]


Deep Ensemble Uncertainty Fails as Network Width Increases: Why, and How to Fix It. Lisa Schut, Edward Hu, Greg Yang, Yarin Gal. ICML Workshop on Uncertainty in Deep Learning (UDL) 2021. [Workshop Paper]

Uncertainty-Aware Counterfactual Explanations for Medical Diagnosis. Lisa Schut*, Oscar Key*, Rory McGrath, Luca Costabello, Bogdan Sacaleanu, Medb Corcoran, Yarin Gal. NeurIPS Workshop on Machine Learning for Health 2020.

[Poster Link]

Capsule Networks -- A Probabilistic Perspective. Lewis Smith, Lisa Schut, Yarin Gal and Mark van der Wilk.

Spotlight paper at the ICML Workshop on Object-Oriented Learning: Perception, Representation, and Reasoning, 2020.

[Workshop Paper] [Full paper] [Lewis' Blogpost]

Influence of Outliers on Cluster Correspondence Analysis. Michel van de Velden, Alfonso Iodice D'Enza, Lisa Schut. CLADAG, Cassino Italy 2019.

[Conference proceedings]

Student Publications

Below are the workshop papers/publications of MSc students I've (co-)supervised:

DeDUCE: Generating Counterfactual Explanations at Scale. Benedikt Höltgen, Lisa Schut, Jan M.Brauner, Yarin Gal. NeurIPS XAI4Debugging Workshop 2021.

Can Network Flatness Explain the Training Speed-Generalisation Connection? Albert Qiaochu Jiang, Clare Lyle, Lisa Schut, Yarin Gal. NeurIPS BDL Workshop 2021.

Invited Talks

  • Counterfactual Explanations: Making AI Decisions More Useful and Trustworthy. Accenture Turing Innovation Symposium 2020

Reviewing Experience


  • ICLR 2021

  • NeurIPS 2021

  • AISTATS 2021

  • ICML 2021


I generated the images on this website using, which uses this BigGAN.