Keynote announcements

[Nikita Kazeev]

Dear multiscale scholars,

It's 2.5 weeks until the deadline - perfect time start finish writing your paper, pack the code and make a submission! I wish I could reward early submissions with easier reviewers, but alas.

Meanwhile, I'm happy to announce the three first keynotes:

Qianxiao Li
Constructing macroscopic dynamics using deep learning
We discuss some recent work on constructing stable and interpretable macroscopic dynamics from trajectory data using deep learning. We adopt a modelling approach: instead of generic neural networks as functional approximators, we use a model-based ansatz for the dynamics following a suitable generalisation of the classical Onsager principle for non-equilibrium systems. This allows the construction of macroscopic dynamics that are physically motivated and can be readily used for subsequent analysis and control. We discuss applications in the analysis of polymer stretching in elongational flow. Moreover, we will also discuss some algorithmic challenges associated with learning (macroscopic) dynamics for scientific applications.

Sergei Gukov
In this talk, we explore the transformative potential of custom reinforcement learning (RL) algorithms in accelerating solutions to complex, research-level mathematical challenges. We begin by illustrating how these algorithms have achieved a 10X improvement in areas where previous advances of the same magnitude required many decades. A comparative analysis of different network architectures is presented to highlight their performance in this context. We then delve into the application of RL algorithms to exceptionally demanding tasks, such as those posed by the Millennium
 Prize problems and the smooth Poincaré conjecture in four dimensions. Drawing on our experiences, we discuss the prerequisites for developing new RL algorithms and architectures that are tailored to these high-level challenges. Based on recent work What makes math problems hard for reinforcement learning: a case study

Charlotte Bunne
Virtual Cells and Digital Twins: Multi-Scale and Multi-Modal AI for Biomedicine
The complexity of cancer demands understanding biological processes across scales, from molecular interactions to tissue architecture. This talk explores how artificial intelligence enables the creation of digital twins at both cellular and tissue levels, with the aim to predict cellular phenotypes, function and their responses to perturbations such as cancer therapies.
Concretely, I will introduce the Virtual Tissues (VirTues) platform, a foundation model framework that transforms how we analyze multiplexed tissue data and seamlessly integrates molecular, cellular, and tissue-scale information to increase diagnostic precision and biological understanding in personalized oncology. VirTues employs a multi-modal vision transformer architecture designed to learn from heterogeneous, high-dimensional datasets spanning different biological markers, measurement characteristics, and variable clinical annotations. While existing approaches often focus on H&E-stained slides, our framework incorporates highly multiplexed imaging techniques that capture hundreds of proteins within single tissue sections. Through unsupervised learning and a multi-scale neural network architecture, VirTues unifies these diverse data sources into a coherent virtual tissue space. As a result, new patient biopsy samples can be automatically mapped into this common representation. This enables integrative analyses of morphological, molecular and spatial complexity while facilitating clinically relevant predictions.
To bridge insights from the analysis of patient samples with personalized treatment, we employ generative models trained on large biomedical datasets. These models predict treatment responses of biopsied cells from metastatic melanoma patients by revealing patterns of signaling pathway modulation associated with driver mutations and metastasis sites. Together, these approaches enable a multi-scale understanding of cancer biology and treatment response, advancing the development of personalized therapies guided by comprehensive digital twins of patients.

Best,
Nikita on behalf of the workshop organisers