CHAI Seminar series
We are excited to bring you the CHAI Seminar Series, organised by our Early Career Researchers (ECRs). This seminar series will provide a platform for knowledge exchange and discussion on cutting-edge research in causal AI and related fields.
Upcoming Seminars
Clinically Grounded and Explainable AI for Chest X-ray Interpretation
Join us for the next CHAI Seminar Series featuring Dr Ajitha Rajan.
Date: Wednesday, 11 November
Time: 1:00 PM (UK time)
While models for automated chest X-ray interpretation can accurately detect a range of pathologies, it can be difficult to establish whether their predictions are based on clinically relevant findings which would be consistent with a human radiologist’s knowledge and expertise.
This talk introduces XpertCausal, a concept-based framework which incorporates radiologist-defined relationships between pathologies and imaging findings (“concepts”) into the design and evaluation of a probabilistic model. Building on our initial work, XpertXAI, which predicts pathology labels using radiologist-defined clinical concepts, XpertCausal also models how pathologies can produce these findings, using a causal structure which is directly informed by clinical domain knowledge.
We will discuss XpertCausal’s performance, explanation quality and limitations when trained and evaluated on public datasets, before outlining our plans for further evaluation using a large private chest X-ray dataset collected in collaboration with Public Health Scotland.
Past Seminars
2026 Seminars
-
Our October Seminar was led by Dr David Watson, Senior Lecturer in Artificial Intelligence at King's College London and Senior Research Scientist at Google DeepMind. In this session, Dr Watson introduced a novel approach for obtaining partial identification in linear systems using leaky instruments, which are permitted to violate the exclusion criterion to a limited extent.
-
Our September Seminar was presented by Dr Jason Hartford (University of Manchester). In this session, Dr Hartford shared his latest research on machine learning, probabilistic modelling, and decision-making under uncertainty, highlighting methods for developing intelligent systems that can make robust decisions in complex real-world environments.
-
Our June Seminar. was presented by Dr. Fani Deligianni. In this session, Dr. Deligianni presented recent advances in physiologically-informed AI methods for high-frequency biosignal analysis and their application in critical care research.
-
Our May Seminar Series featured two complementary talks exploring innovative approaches to the early detection and characterisation of Alzheimer’s disease. Together, they brought perspectives from neuroscience, biomedical engineering and network modelling.
Talk 1 : From Eye to Brain: Investigating EEG and Retinal Measures for Early Detection of Alzheimer’s Disease by Melody Sequeira
Talk 2 : Granular Changes in Brain Networks: Multiplex Nodal Modularity for Alzheimer’s Disease Characterisation and Simulation by Avalon Campbell‑Cousins.
-
Our April Seminar Series was led by Professor Francesco Rubino, a Bariatric Surgeon at KCL.
Unfortunately, due to technical difficulties we were unable to record this session.
-
Our March seminar “Optimal decision rules with policy-relevant guarantees” was led by Dr. Mats Julius Stensrud, Chair of Biostatistics at EPFL in Lausanne.
-
Our February seminar “Assessing Real-World Performance of AI-Based Neuroimaging Diagnosis Tools Through Hospital Data Warehouses”was delievered by Ninon Burgos who is a CNRS research director at the Paris Brain Institute in France, co-head of the ARAMIS Lab, and PR[AI]RIE fellow.
-
Our January seminar “Causal Effect Estimation with Context and Confounders”was led by Professor Arthur Gretton, UCL.
2025 Seminars
-
Our December seminar “Towards Inherently Interpretable Neural Networks for Differential Diagnosis” was led by Tom Nuno Wolf, a PhD candidate at TU Munich.
-
Our November seminar “Amiguous Interventions and Causal Abstractions” was led by Dr. Sander Beckers, University College London (UCL).
-
Our October seminar “Causal Roadmap for generation of reliable Real-World Evidence with application to genomics and biomarker discovery” will be led by Dr. Ava Khamesh, University of Edinburgh.
-
Our September seminar “See, Think, Act: Training Agents by Reinforcing Reasoning” was led by Dr. Maning Li, Northwestern University.
-
Our June seminar “How Causal Analysis Can Help Improve Radiation Cancer Treatment” was led by Dr. Eliana Vasquez Osorio, University of Manchester.
-
Our May seminar titled “On the Identifiability of Switching Dynamical Systems”, was given by Dr Yingzhen Li, Imperial College London.
-
Our April seminar titled “Regulating AI for Healthcare: Developments, Challenges, and Prospects.”, was given by James Lowe, University of Exeter.