Bio

Philipp Frank is an Astronomy and Machine Learning researcher who is developing and applying statistical and ai methods to help deepen our understanding of the structure of the Milky Way and the Cosmos. He did his PhD and a followup Postdoc in Germany at Ludwig Maximilians University and the Max-Planck-Institute for Astrophysics where he worked on probabilistic ML and numerical inference methods and contributed to applications ranging from radio interferometry, X- and gamma-ray imaging, Cosmic Ray air-shower reconstructions, and 3d maps of the dust and gas content of our local Galactic neighborhood.
As a KIPAC Fellow at Stanford he aims to push 3D mapping of the interstellar medium to unprecedented scales in both size and resolution, and incorporate multiple additional tracers for a more comprehensive picture of local structures. This aims to shed light on the mechanisms of star formation and galaxy dynamics across scales only accessible through our unique vantage point within the Galaxy.

Stanford Advisors

Contact

  • Academic
    phfrank@stanford.edu
    University - Scholar Department: Physics Position: Postdoctoral Scholar

Additional Info

Links

All Publications

  • Spatially coherent 3D distributions of HI and CO in the Milky Way ASTRONOMY & ASTROPHYSICS Soeding, L., Edenhofer, G., Ensslin, T. A., Frank, P., Kissmann, R., Phan, V., Ramirez, A., Zandinejad, H., Mertsch, P. 2025; 693

    View details for DOI 10.1051/0004-6361/202451361

    View details for Web of Science ID 001397121400010

  • Towards a Field-Based Bayesian Evidence Inference from Nested Sampling Data ENTROPY Westerkamp, M., Roth, J., Frank, P., Handley, W., Ensslin, T. 2024; 26 (11)

    Abstract

    Nested sampling (NS) is a stochastic method for computing the log-evidence of a Bayesian problem. It relies on stochastic estimates of prior volumes enclosed by likelihood contours, which limits the accuracy of the log-evidence calculation. We propose to transform the prior volume estimation into a Bayesian inference problem, which allows us to incorporate a smoothness assumption for likelihood-prior-volume relations. As a result, we aim to increase the accuracy of the volume estimates and thus improve the overall log-evidence calculation using NS. The method presented works as a post-processing step for NS and provides posterior samples of the likelihood-prior-volume relation, from which the log-evidence can be calculated. We demonstrate an implementation of the algorithm and compare its results with plain NS on two synthetic datasets for which the underlying evidence is known. We find a significant improvement in accuracy for runs with less than one hundred active samples in NS but a proneness for numerical problems beyond this point.

    View details for DOI 10.3390/e26110930

    View details for Web of Science ID 001364752400001

    View details for PubMedID 39593875

    View details for PubMedCentralID PMC11592769

  • Disentangling the Faraday rotation sky ASTRONOMY & ASTROPHYSICS Hutschenreuter, S., Haverkorn, M., Frank, P., Raycheva, N. C., Ensslin, T. A. 2024; 690

    View details for DOI 10.1051/0004-6361/202346740

    View details for Web of Science ID 001339205700009

https://orcid.org/0000-0001-5610-3779