Abstract

The Cancer Research UK Cambridge Centre and the Department of Radiology at Addenbrooke’s are pleased to announce a seminar series on Artificial Intelligence (AI) in Medicine, which aims to provide a comprehensive overview of the latest developments in this rapidly evolving field. As AI continues to revolutionize healthcare, we believe it is essential to explore its potential and discuss the challenges and opportunities it presents.

The seminar series will feature prominent experts in the field who will share their research and insights on a range of topics, including AI applications in disease diagnosis, drug discovery, and patient care. Each seminar will involve two talks, followed by an interactive discussion with a light lunch! We hope that this seminar series will be a valuable platform for researchers, practitioners and students to learn about the latest trends and explore collaborations in the exciting field of AI in Medicine.

The next seminar will be held on 30 January 2024, 12-1pm at the Jeffrey Cheah Biomedical Centre (Main Lecture Theatre), University of Cambridge and streamed online via Zoom. This month will feature the following two talks:

Multimodal, data-efficient, and robust AI for real-world biosignals and the role of generative models – Dr Dimitris Spathis, Senior Research Scientist at Nokia Bell Labs and Visiting Researcher at University of Cambridge

Dimitris Spathis is a senior research scientist at Nokia Bell Labs and a visiting researcher at the University of Cambridge, where he completed his PhD. His research enables machine learning to handle complex real-world data efficiently, with a particular interest in health sensing. He has previously worked at Microsoft Research, Telefonica, and Ocado. He serves on the program committees of top AI conferences such as AAAI , IJCAI, and KDD , and the editorial board of Nature Digital Medicine. Read more: https://dispathis.com/.

Abstract: The limited availability of labels for machine learning on multimodal data hampers progress in the field. In this talk, I will discuss our recent efforts to address this problem, building on the paradigms of self-supervised and multimodal learning. With models such as CroSSL, Step2Heart, and SelfHAR, we put forward principled ways to learn generalizable representations from high-resolution data through masking, knowledge distillation, and physiology-inspired pre-training. We show that these models can be applied to various clinically relevant applications to improve mental health, fitness, sleep, and voice-based diagnostics. At the same time, due to data size limitations, these models are limited in size and generalization capabilities compared to popular generative models such as GPT . What if we could use Large Language Models (LLMs) as data-agnostic pre-trained models? I will close the talk by highlighting where LLMs fail in processing sequential data as text tokens and some ideas on how to address the critical “modality gap”.

Using machine learning methods to improve classification and prediction of psychiatric conditions – Dr Katharina Zühlsdorff, Visiting Postdoctoral Fellow at Department of Psychology

Katharina Zühlsdorff is a graduate-entry Medicine student and researcher in Cognitive Computational Neuroscience at the University of Cambridge. She is a Foulkes Foundation Fellow (2023 intake) and Downing College Bye-Fellow in Psychology. She completed her PhD at the Department of Psychology, University of Cambridge and the Alan Turing Institute, London, in September 2022. She researches the behavioural and neural circuits underlying reinforcement learning in neuropsychiatric disorders such as substance use disorder and depression. Moreover, she works on the application of deep learning algorithms to large-scale, multimodal datasets to identify patterns that may help us understand the aetiology of psychiatric diseases better.

Abstract: Cognitive flexibility can be investigated using tests such as probabilistic reversal learning (PRL). In various neuropsychiatric conditions, including substance use disorders, gambling disorder, major depressive disorder and schizophrenia, overall impairments in PRL flexibility are observed. Using reinforcement learning (RL) models, a deeper mechanistic explanation of the latent processes underlying flexibility can be gained. I will present results from an analysis of PRL data from individuals with different psychiatric diagnoses using a hierarchical Bayesian RL approach and relate behavioural findings to the underlying neural substrates. Furthermore, I will discuss how graph neural network models can be used to incorporate cognitive and neuroimaging data to improve prediction of psychiatric conditions.

This is a hybrid event so you can also join via Zoom:

https://zoom.us/j/99050467573?pwd=UE5OdFdTSFdZeUtIcU1DbXpmdlNGZz09

Meeting ID: 990 5046 7573 and Passcode: 617729

We look forward to your participation! If you are interested in getting involved and presenting your work, please email Ines Machado at im549@cam.ac.uk