Prior-Fitted Functional Flows: In-Context Generative Models for Pharmacokinetics

César Ojeda, Niklas Hartung, Purity Kamene Kavwele, Tim Jahn, Piyush Kumar, Marian Klose, Wilhelm Huisinga, Ramsés J Sánchez, Darius A Faroughy
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:5040-5059, 2026.

Abstract

We introduce Prior-Fitted Functional Flows, a generative foundation model for pharmacokinetics that enables zero-shot population synthesis and individual forecasting without manual parameter tuning. We learn functional vector fields, explicitly conditioned on the sparse, irregular data of an entire study population. This enables the generation of coherent virtual cohorts as well as forecasting of partially observed patient trajectories with calibrated uncertainty. We construct a new open-access literature corpus to validate our priors, and demonstrate state-of-the-art predictive accuracy on extensive real-world datasets. Our code repository a, pretrained model, data and examples are available online.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-ojeda26a, title = {Prior-Fitted Functional Flows: In-Context Generative Models for Pharmacokinetics}, author = {Ojeda, C\'{e}sar and Hartung, Niklas and Kavwele, Purity Kamene and Jahn, Tim and Kumar, Piyush and Klose, Marian and Huisinga, Wilhelm and S\'{a}nchez, Rams\'{e}s J and Faroughy, Darius A}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {5040--5059}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/ojeda26a/ojeda26a.pdf}, url = {https://proceedings.mlr.press/v337/ojeda26a.html}, abstract = {We introduce Prior-Fitted Functional Flows, a generative foundation model for pharmacokinetics that enables zero-shot population synthesis and individual forecasting without manual parameter tuning. We learn functional vector fields, explicitly conditioned on the sparse, irregular data of an entire study population. This enables the generation of coherent virtual cohorts as well as forecasting of partially observed patient trajectories with calibrated uncertainty. We construct a new open-access literature corpus to validate our priors, and demonstrate state-of-the-art predictive accuracy on extensive real-world datasets. Our code repository a, pretrained model, data and examples are available online.} }
Endnote
%0 Conference Paper %T Prior-Fitted Functional Flows: In-Context Generative Models for Pharmacokinetics %A César Ojeda %A Niklas Hartung %A Purity Kamene Kavwele %A Tim Jahn %A Piyush Kumar %A Marian Klose %A Wilhelm Huisinga %A Ramsés J Sánchez %A Darius A Faroughy %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-ojeda26a %I PMLR %P 5040--5059 %U https://proceedings.mlr.press/v337/ojeda26a.html %V 337 %X We introduce Prior-Fitted Functional Flows, a generative foundation model for pharmacokinetics that enables zero-shot population synthesis and individual forecasting without manual parameter tuning. We learn functional vector fields, explicitly conditioned on the sparse, irregular data of an entire study population. This enables the generation of coherent virtual cohorts as well as forecasting of partially observed patient trajectories with calibrated uncertainty. We construct a new open-access literature corpus to validate our priors, and demonstrate state-of-the-art predictive accuracy on extensive real-world datasets. Our code repository a, pretrained model, data and examples are available online.
APA
Ojeda, C., Hartung, N., Kavwele, P.K., Jahn, T., Kumar, P., Klose, M., Huisinga, W., Sánchez, R.J. & Faroughy, D.A.. (2026). Prior-Fitted Functional Flows: In-Context Generative Models for Pharmacokinetics. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:5040-5059 Available from https://proceedings.mlr.press/v337/ojeda26a.html.

Related Material