CMIC/WEISS Joint Seminar Series

First Talk:   

Title: Increased Belief Instability in Psychotic Disorders Predicts Treatment Response to Metacognitive Training - Daniel Hauke

Abstract: Background and Hypothesis: 
In a complex world, gathering information and adjusting our beliefs about the world is of paramount importance. The literature suggests that patients with psychotic disorders display a tendency to draw early conclusions based on limited evidence, referred to as the jumping-to-conclusions bias, but few studies have examined the computational mechanisms underlying this and related belief-updating biases. Here, we employ a computational approach to understand the relationship between jumping-to-conclusions, psychotic disorders, and delusions. 

Bio: Daniel Hauke After a bachelor’s in psychology at the Georg-August-University Goettingen in Germany and the Universidade Federal do Ceará in Brazil, Daniel obtained a Master in Cognitive Neuroscience following studies at Maastricht University and the Translational Neuromodeling Unit, University of Zurich and ETH Zurich. During his PhD in Computer Science, he cast different symptoms of schizophrenia as instances of hierarchical Bayesian inference and used machine learning to predict clinically relevant outcomes such as treatment response and transition to psychosis. Since December 2022, he has joined Rick Adam’s team at the CMIC at UCL and focusses on developing biophysically informed models of EEG data to measure neuroreceptor function in patients with schizophrenia non-invasively. Other research interests of Daniel’s are extending these approaches to other psychiatric conditions and understanding the effects of psychedelics on the brain. 

Second Talk

Title: Scalable Bayesian uncertainty quantification with learned convex regularisers - Tobias Liaudat

Abstract: The last decade brought us substantial progress in computational imaging techniques for current and next-generation interferometric telescopes, such as the SKA. Imaging methods have exploited sparsity and more recent deep learning architectures with remarkable results.  Despite good reconstruction quality, obtaining reliable uncertainty quantification (UQ) remains a common pitfall of most imaging methods. The UQ problem can be addressed by reformulating the inverse problem in the Bayesian framework. The posterior probability density function provides a comprehensive understanding of the uncertainties. However, computing the posterior in high-dimensional settings is an extremely challenging task. Posterior probabilities are often computed with sampling techniques, but these cannot yet cope with the high-dimensional settings from radio imaging. This work proposes a method to address uncertainty quantification in radio-interferometric imaging with data-driven (learned) priors for very high-dimensional settings. Our model uses an analytic physically motivated model for the likelihood and exploits a data-driven prior learned from data. The proposed prior can encode complex information learned implicitly from training data and improves results from handcrafted priors (e.g., wavelet-based sparsity-promoting priors). We exploit recent advances in neural-network-based convex regularisers for the prior that allow us to ensure the log-concavity of the posterior while still being expressive. We leverage probability concentration phenomena of log-concave posterior functions that let us obtain information about the posterior avoiding the use of sampling techniques. Our method only requires the maximum-a-posteriori (MAP) estimation and evaluations of the likelihood and prior potentials. We rely on convex optimisation methods to compute the MAP estimation, which are known to be much faster and better scale with dimension than sampling strategies. The proposed method allows us to compute local credible intervals, i.e., Bayesian error bars, and perform hypothesis testing of structure on the reconstructed image. We demonstrate our method by reconstructing simulated radio-interferometric images and carrying out fast and scalable uncertainty quantification.  

Bio: Tobias Liaudat Tobias is a postdoctoral fellow working with Professors Marta Betcke from UCL’s Computer Science department, Jason McEwen from UCL’s MSSL department, and Marcelo Pereyra from Heriot-Watt University in Edinburgh. Previously he did his PhD at the CEA research centre in Saclay, France and the Université Paris-Saclay working with Drs Jean-Luc Starck and Martin Kilbinger. Tobias' research is focused on inverse problems in imaging sciences for different applications in astrostatistics like point spread function modelling and radio imaging reconstruction. He is developing new methodology combining applied mathematics, machine learning, and physics from the inverse problem.  

Chair : Laura Panagiotaki

Link to the Moodle page here: https://moodle.ucl.ac.uk/course/view.php?id=19613   Please enrol with key: CMIC  Alternatively you can use the zoom link below to go direct to the seminar  https://ucl.zoom.us/j/99464005163?pwd=ZFdURkJ4TjJIeGVhbXpTclhuNE9WUT09