Schedule

Title: We need to talk about AI

Speaker:  Eva Navarro López

Abstract: Artificial Intelligence (AI) is a term that has been widely misused in recent years. In most cases, AI-related technologies do not imply any form of intelligence. The true goal of AI is to model and formalise aspects of intelligence –in humans and in nature– using computational and mathematical paradigms. However, attempting to reproduce or simulate intelligence does not mean that the resulting systems are equipped with intelligence themselves.

Data science, which may use techniques from AI, comprises any form of computation that involves data and extends far beyond traditional statistics. Today, many techniques labelled as “AI” are primarily data-driven, which are just a small part of what AI is. Data-driven AI relies on the acquisition, classification and analysis of large amounts of data. But data-driven and statistical methods alone are insufficient for understanding complex dynamical systems. These systems, characterised by emergent and collective behaviours, continuous dynamic evolution, and intricate interactions with their environment and internal components, defy purely statistical or data-driven approaches. They are inherently unpredictable and require adaptive models that account for intrinsic discontinuities.

To model and analyse complex dynamical systems, we must move beyond ‘only discrete’ models, conventional statistical paradigms and data-driven techniques, which stem from the orthodoxy of computational theory and machine learning.

In this talk, I will present key findings from my research on the modelling, analysis, and control of complex dynamical systems. I will trace the natural evolution of my work, transitioning from control systems to self-organising adaptive networks of networks –including the human brain– within the broader scope of nature-inspired machine intelligence and artificial intelligence. Along this journey, I will explore the pivotal area of hybrid dynamical systems and cyber-physical systems, touching on formal specifications of dynamical properties, symbolic AI and automated verification methods –each guided by the beauty and elegance of their mathematical foundations. 

Inspired by the working of the human brain, self-organisation in nature and Alan Turing’s morphogenesis –having collaborated with Alan Turing’s last student– this talk will culminate in a reflection on the current AI status quo and my work on social computing and AI activism. 

Title: Bayesian graph-informed disease modeling from neuroimaging to EHR data

Speaker: Yize Zhao

Abstract: Neurodegenerative and complex chronic diseases emerge from interactions that span biological networks and populations. I will introduce two Bayesian frameworks that quantify these interactions, from within-brain propagation to across-disease comorbidity. In the first project, aimed at characterizing Tau protein spread along functional networks in the early course of Alzheimer’s disease generated from the A4 study, we jointly model tau propagation, functional connectivity structure, and subgroup heterogeneity using cross-sectional data. By integrating graph-constrained infection dynamics with connectivity patterns, the model infers plausible propagation pathways and subgroup-specific infection sequences. In the second project, using UK Biobank EHRs, we represent disease relationships through a latent hypergraph, where each hyperedge captures the higher-order clusters of diseases that share covariate-dependent risk. By uncovering disease hyperedges and their associated risk factors, we characterize how biological and lifestyle factors jointly influence sets of related conditions. To scale posterior inference while preserving uncertainty, we employ amortized variational inference with neural parameterization. Together, these studies yield probabilistic, graph-informed and interpretable views of disease pathology and organization.

Title: Learning in the space of probability measures

Speaker: Caroline Moosmueller

Abstract: Many datasets in modern applications, from cell gene expression and images to shapes and text documents, are naturally interpreted as probability measures, distributions, histograms, or point clouds. This perspective motivates the development of learning algorithms that operate directly in the space of probability measures. However, this space presents unique challenges: it is nonlinear and infinite-dimensional. Fortunately, it possesses a natural Riemannian-type geometry which enables meaningful learning algorithms. 

This talk will provide an introduction to the space of probability measures and present approaches to unsupervised, supervised, and manifold learning within this framework. We will examine temporal evolutions on this space, including flows involving stochastic gradient descent and trajectory inference, with applications to analyzing gene expression in single cells. The proposed algorithms are furthermore demonstrated in pattern recognition tasks in imaging and medical applications.

Title: Scientific Machine Learning Enhanced Forecasting in Subsurface Flows

Speaker: Hannah Lu

Abstract: Accurate forecasting of subsurface flow migration and trapping remains a central challenge in geoscience. Despite well-established physical understanding and advanced numerical simulators, predictive uncertainty persists due to geological heterogeneity, data scarcity, and the prohibitive cost of high-fidelity simulations. This talk presents a scientific machine learning  framework to enhance forecasting capabilities for subsurface flows across scales. Starting from lab-scale experiments, we develop digital twins that combine experimental data, physics-based models, and machine learning surrogates to reproduce and predict multiphase flow dynamics under both pressure-driven and capillary–buoyancy–dominated flow regimes. The framework enables quantitative studies of energy storage retention and flow transport, while identifying key geological and barrier properties that control storage/operation performance. Through collaborations, this research aims to establish a scalable and interpretable SciML foundation for digital twins of subsurface energy storage, bridging experimental observables with predictive modeling of subsurface flow in realistic heterogeneous formations.

Data Science for the Biological Sciences - Workshop 1

Speakers: Kathleen Lois Foster & Alessandro Maria Selvitella

Abstract: This Workshop is aimed at students and practitioners in the biological sciences who are interested in developing coding and data science skills to solve concrete problems emerging in their biological field of study. By the end of the workshops, the participants will have gathered technical and theoretical skills in data science. They will have learned how to install R and R-studio, use the statistical software R to perform basic statistical analysis of a biological question, and visualize the biological information hidden in the data under study. Furthermore, they will have gained knowledge about the structure of different types of data, descriptive statistics, including mean, standard deviation, confidence intervals, and probability distributions, how to perform hypothesis tests, including t-test and ANOVA, and the difference between statistical and biological significance.

Slides

R-code

[In person only or specific request must be sent to the organizers for online participation]

Data-Driven Dynamical Systems - Workshop 1

Speakers: Alessandro Maria Selvitella

Abstract: This Workshop is aimed at graduate students with some background in differential equations and machine learning. The workshop will cover topics such as Runge-Kutta methods, Physics Informed Neural Networks, and System Identification of Nonlinear Dynamics. These tools will be illustrated on a toy problem, a dynamical system modeling the central pattern generator of the lamprey.

[In person only or specific request must be sent to the organizers for online participation]

Title: Statistical Inference for Networks of Point Processes

Speaker: Ali Shojaie

Abstract: Advances in calcium florescent imaging and Neuropixel assays have facilitated monitoring of the activity of thousands of neurons in live animals. Data from these live images reveal the firing times of neurons, or their spike trains, in response to various stimuli. In this talk, we will discuss new methodological, computational and theoretical developments for learning functional connectivity networks from high-dimensional Hawkes processes, which are widely used to model neuronal spike train data. In particular, we will discuss new procedures for learning connectivity networks from high-dimensional point processes and statistical inference procedures for characterizing the uncertainty of the resulting estimators. We also discuss an extension of this procedure to learn networks from multiple experiments, which are commonly used to glean insight into changes in brain connectivity associated with different tasks, as well as new developments for spectral analysis of high-dimensional point processes.

Data Science for the Biological Sciences - Workshop 2

Speakers: Kathleen Lois Foster & Alessandro Maria Selvitella

Abstract: This Workshop is aimed at students and practitioners in the biological sciences who are interested in developing coding and data science skills to solve concrete problems emerging in their biological field of study. By the end of the workshops, the participants will have gathered technical and theoretical skills in data science. They will have learned how to install R and R-studio, use the statistical software R to perform basic statistical analysis of a biological question, and visualize the biological information hidden in the data under study. Furthermore, they will have gained knowledge about the structure of different types of data, descriptive statistics, including mean, standard deviation, confidence intervals, and probability distributions, how to perform hypothesis tests, including t-test and ANOVA, and the difference between statistical and biological significance.

Slides

R-code

[In person only or specific request must be sent to the organizers for online participation]

Data-Driven Dynamical Systems - Workshop 2

Speakers: Alessandro Maria Selvitella

Abstract: This Workshop is aimed at graduate students with some background in differential equations and machine learning. The workshop will cover topics such as Runge-Kutta methods, Physics Informed Neural Networks, and System Identification of Nonlinear Dynamics. These tools will be illustrated on a toy problem, a dynamical system modeling the central pattern generator of the lamprey.

[In person only or specific request must be sent to the organizers for online participation]