The Joint Statistical Meetings bring together the largest annual gathering of statisticians in the world. The conversation at JSM reflects how deeply statistics and machine learning are converging in applied science. For drug development, that convergence is particularly consequential: the tools emerging from generative AI and deep learning are now meeting the rigor of mechanistic, population-level modeling. At JSM 2026 in Boston, our team will present two talks exploring exactly that frontier, showing how nonlinear mixed-effects models can be extended with modern generative and multimodal AI to extract more from the data we already have.

Date: August 1 - 6, 2026

Location: Boston, Massachusetts

Podium

Topic: A Deep Dive into Generative Scientific Machine Learning

Date: August 6, 2026 | 11:35 am - 11:55 am

Speaker: Lorenzo Contento

This talk examines how combining mechanistic modeling with generative ML produces models that are both interpretable and flexible without sacrificing the generalization you need when datasets are small. Lorenzo will show how NLME models, understood as generative models over subject populations, connect naturally to variational autoencoders and other generative architectures, and how that parallel opens up a rich set of tools for pharmacometrics. Know more

Topic: Multimodal modelling for bridging information across datasets in drug development

Date: August 6, 2026 | 11:55 am - 12:15 pm

Speaker: Niklas Korsbo

Patient data is increasingly complex — images, clinical notes, longitudinal measurements, but conventional NLME models have struggled to incorporate it. Niklas will present a conceptual framework for doing exactly that, using pre-trained embedding and generative models to transform complex data modalities into representations that integrate cleanly with NLME. The result is a coherent approach to multimodal pharmacometric modeling that opens up dramatic improvement of our ability to leverage the information we have about our patients. Know more

Whether you are rethinking how mechanistic models handle complex patient data, exploring where generative AI fits in your pharmacometrics workflow, or simply curious about what SciML looks like in practice, meet our team in Boston.