About

I am a Physicist and Postdoctoral researcher at the Max-Planck-Institute for Astrophysics (MPA) with a primary reserach interest in information theory, machine learning, and observational astronomy. I did my PhD in the Information Field Theory group of Torsten Enßlin at MPA. My work focused on the development of theoretical models for probabilistic inference as well as models for fast numerical approximation. The algorithmic realizations of my work can largely be found within the software package Numerical Information Field Theory (NIFTy), a python library to perform Bayesian inference for field like quantities, which I am actively developing at MPA. Before the PhD, I did my Masters at the Faculty of Physics at Ludwig-Maximilians-Universität in Munich working on statistical modelling of dynamical systems.

Publications

See also: Google Scholar

First author

• Geometric Variational Inference and Its Application to Bayesian Imaging; Frank, P.; Phys. Sci. Forum 2022, 5(1), 6. doi.
• Variable structures in M87* from space, time and frequency resolved interferometry; Arras*, P., Frank*, P., Haim*, P., Knollmüller*, J., Leike*, R., Reinecke, M., Enßlin, T.A.; Nature Astron (2022) (*shared first authors). arXiv, doi, code, data products.
• Geometric variational inference; Frank, P., Leike, R., and Enßlin, T.A.; Entropy 2021, 23, 853. arXiv, doi, cover story.
• Field dynamics inference for local and causal interactions; Frank, P., Leike, R., and Enßlin, T.A.; Annalen der Physik 2021, 533, 2000486. arXiv, doi.
• Field dynamics inference via spectral density estimation; Frank, P., Steininger, T., and Enßlin, T.A.; Physical Review E 2017, 96, 052104. arXiv, doi.
• SOMBI: Bayesian identification of parameter relations in unstructured cosmological data; Frank, P., Jasche, J., and Enßlin, T.A.; Astronomy & Astrophysics 2016, 595, A75. arXiv, doi.

Preprints

• Probabilistic simulation of partial differential equations; Frank, P., Enßlin, T.A.; ArXiv preprints 2020, arXiv:2010.06583. arXiv.

Co-author

• Butterfly Transforms for Efficient Representation of Spatially Variant Point Spread Functions in Bayesian Imaging; Eberle, V.; Frank, P.; Stadler, J.; Streit, S.; Enßlin, T.A.; Entropy 2023, 25(4), 652. doi.
• Analysis of Dynamical Field Inference in a Supersymmetric Theory; Westerkamp, M.; Ovchinnikov, I.; Frank, P; Enßlin, T.A.; Phys. Sci. Forum 2022, 5(1), 27. doi.
• Towards Moment-Constrained Causal Modeling; Guardiani, M., Frank, P., Kostić, A., Enßlin, T.A.; Phys. Sci. Forum 2022, 5(1), 7. doi.
• Efficient Representations of Spatially Variant Point Spread Functions with Butterfly Transforms in Bayesian Imaging Algorithms; Eberle, V.; Frank, P.; Stadler, J.; Streit, S.; Enßlin, T.A.; Phys. Sci. Forum 2022, 5(1), 33. doi.
• Dynamical field inference and supersymmetry; Westerkamp, M., Ovchinnikov, I., Frank, P., Enßlin, T.A.; Entropy 2021, 23, 1652. arXiv, doi.
• Probabilistic Autoencoder using Fisher Information; Zacherl, J., Frank, P., and Enßlin, T.A.; Entropy 2021, 23, 1640. arXiv, doi.
• Bayesian decomposition of the Galactic multi-frequency sky using probabilistic autoencoders; Milosevic, S., Frank, P., Leike, R., Müller, A., Enßlin, T.A.; Astronomy & Astrophysics 2021, 650, A100. arXiv, doi, data products.
• Reconstructing non-repeating radio pulses with Information Field Theory; Welling, C., Frank, P., Enßlin, T.A., and Nelles, A.; Journal of Cosmology and Astroparticle Physics 2021, 04, 071. arXiv, doi.
• Non-parametric Bayesian Causal Modeling of the SARS-CoV-2 Viral Load Distribution vs. Patient's Age; Guardiani, M., Frank, P., Kostić, A., Edenhofer, G., Roth, J., Uhlmann, B., Enßlin, T.A.; PLoS ONE 17(10): e0275011. arXiv, doi
• Unified radio interferometric calibration and imaging with joint uncertainty quantification; Arras, P., Frank, P., Leike, R., Westermann, R., Enßlin, T.A.; Astronomy & Astrophysics 2019, 627, A134. arXiv, doi.
• Nifty5: Numerical information field theory v5; Arras, P., Baltac, M., Enßlin, T.A., Frank, P., Hutschenreuter, S., Knollmüller, J., Leike, R., Newrzella, M., Platz, L., Reinecke, M., Stadler, J.; Astrophysics Source Code Library 2019, ascl: 1903.008. ascl, repository.
• NIFTy 3 - Numerical Information Field Theory: A Python Framework for Multicomponent Signal Inference on HPC Clusters; Steininger, T., Dixit, J., Frank, P., Greiner, M., Hutschenreuter, S., Knollmüller, J., Leike, R., Porqueres, N., Pumpe, D., Reinecke, M., Šraml, M., Varady, C., Enßlin, T.A.; Annalen der Physik 2019, 531, 1800290. arXiv, doi.
• Sharpening up Galactic all-sky maps with complementary data - A machine learning approach; Müller, A., Hackstein, M., Greiner, M., Frank, P., Bomans, D.J., Dettmar, R., Enßlin, T.A.; Astronomy & Astrophysics 2018, 620, A64. arXiv, doi, data products.
• STARBLADE: STar and Artefact Removal with a Bayesian Lightweight Algorithm from Diffuse Emission; Knollmüller, J., Frank, P., Enßlin, T.A.; Astrophysics Source Code Library 2018, ascl: 1805.009. ascl, repository.

Preprints

• Disentangling the Faraday rotation sky; Hutschenreuter, S; Haverkorn, M; Frank, P; Raycheva, N.C.; Enßlin, T.A.; ArXiv preprints 2023, arXiv:2304.12350. arXiv.
• Measurement in a Unitary World; Johnson, V.; Leike, R.; Frank, P.; Enßlin, T.A.; ArXiv preprints 2022, arXiv:2212.03829. arXiv.
• Sparse Kernel Gaussian Processes through Iterative Charted Refinement (ICR); Edenhofer, G., Leike, R.H., Frank, P., Enßlin, T.A.; ArXiv preprints 2022, arXiv:2206.10634. arXiv.
• The Galactic 3D large-scale dust distribution via Gaussian process regression on spherical coordinates; Leike, R.H., Edenhofer, G., Knollmüller, J., Alig, C., Frank, P., Enßlin, T.A.; ArXiv preprints 2022, arXiv:2204.11715. arXiv, data products.
• Multi-Component Imaging of the Fermi Gamma-ray Sky in the Spatio-spectral Domain; Platz, L.I., Knollmüller, J., Arras, P., Frank, P., Reinecke, M., Jüstel, D., Enßlin, T.A.; ArXiv preprints 2022, arXiv:2204.09360. arXiv, data products.
• Separating diffuse from point-like sources-a Bayesian approach; Knollmüller, J., Frank, P., Enßlin, T.A.; ArXiv preprints 2018, arXiv:1804.05591. arXiv.

Theses

• PhD thesis (Dissertation); Approximate inference in astronomy; 2021; doi.
• Master thesis; Field dynamics inference via spectral density estimation; 2018; pdf.
• Bachelor thesis; Self Organizing Maps and Bayesian Inference in cosmology; 2015; pdf.

Teaching

Co-Supervised Master Students

• David Gorbunov; Density reconstruction using Geometric Variational Inference.
• Johannes Zacherl; Probabilistic Autoencoder using Fisher Information.
• Matteo Guardiani; Non-parametric Bayesian Causal Modeling of the SARS-CoV-2 Viral Load Distribution vs. Patient's Age.
• Vincent Eberle; Efficient representation of Instrument Responses.
• Sara Milosevic; Astrophysical data analysis with variational autoencoders.
• Margret Westerkamp; Dynamical Field Inference via Ghost Fields.
• Morten Giese; Inference of the atmospheric electron density with LOFAR data.

Tutoring

• Tutor and head of the tutoring group for the Information Theory & Information Field Theory lectures; 2017 - 2020; Ludwig-Maximilians-University Munich, Germany.

School teaching

• Math teacher for children with special needs and mentally disabled children at the Augustinum Landschulheim Elkhofen; 02/2022 - 07/2022; Elkhofen, Germany.

Talks, Conferences & Workshops

Invited talks

• VLTI and ALMA Synthesis Imaging Workshop, European Southern Observatory, January 9-12, 2023, Garching, Germany; event, pdf.
• iid2022: Statistical Methods for Event Data, Lake Guntersville State Park, November 15-18, 2022, Guntersville, AL; event, pdf, tutorial, resources.
• Seminar, MIT Haystack Observatory, 09.11.2022, Westford, MA, USA; pdf.
• NIFTy tutorial, MIT Haystack Observatory, 09.11.2022, Westford, MA, USA; pdf, resources .
• NIFTy tutorial, Black Hole Initiative - Harvard University, 08.11.2022, Cambridge, MA, USA; pdf, resources .
• Group Seminar, Black Hole Initiative - Harvard University, 07.11.2022, Cambridge, MA, USA; pdf.
• LMU seminar: Signal reconstruction with Python, Max-Planck-Institute for Astrophysics, 21.09.2022, Garching, Germany; pdf.
• Probabilistic Variational Autoencoders, Berlin Machine Learning Group, 07.02.2022, Berlin, Germany; (remote) Event Webpage, Recording, pdf.
• LMU seminar: Signal reconstruction with Python, Max-Planck-Institute for Astrophysics, 30.09.2021, Garching, Germany; pdf.
• ODSL Journal Club, Excellence Cluster Origins, 16.07.2021, Garching, Germany; pdf.
• Seminar, Center for Astrophysics, 28.06.2021, Cambridge, Massachusetts, United States; (remote) pdf.
• Institute Seminar, Max-Planck-Institute for Astrophysics, 04.05.2020, Garching, Germany; pdf.

Conferences & Workshops

• IMAGINE workshop: Towards a comprehensive model of the galactic magnetic field, Nordita, April 3-28, 2023, Stockholm, Sweden; event, talk, gp session, tutorial.
• Wissenschaftskonferenz: Science from Space, German Aerospace Center (DLR), 27-28 September 2022, Bonn, Germany; event.
• SOcraSCALES Workshop, MIAPbP Munich Institute for Astro-, Particle and BioPhysics, Excellence Cluster ORIGINS, 5-16 September 2022, Garching, Germany; event, pdf.
• 41st MaxEnt2022 Conference, Institute Henri Poincaré, 18-22 July 2022, Paris, France; event, pdf.
• ngEHT Meeting: "Assembling the ngEHT", 22-25 June 2022, Granada, Spain; event, resources, pdf.
• NIFTy tutorial at "Astroparticle School", Erlangen Center for Astroparticle Physics, 2-10 October 2019, Obertrubach-Bärnfels, Germany; event, resources.
• NIFTy tutorial at "Big Data Science in Astroparticle Research", RWTH Aachen University SuperC, 18-20 February 2019, Aachen, Germany; event, resources.
• Bayesian inference meets radio reality, Max-Planck-Institute for Astrophysics, 15-20 July 2018, Garching, Germany; event.
• Interdisciplinary Cluster Workshop "Challenges in statistical inference", Excellence Cluster Universe, 7-9 November 2016, Garching, Germany; event, pdf.

Software

• NIFTy (Numerical Information Field Theory): repository, documentation, pip.
• Dynamic VLBI resolve (Bayesian imaging for dynamic very long basline interferometry): repository.

Data products

• The Galactic Cartography Portal: website.
• Spatio-spectral multi-component maps of the Fermi gamma-ray sky: doi.
• Decomposition of the Galactic multi-frequency sky with autoencoders: doi.
• Time-resolved reconstruction of the Black hole shadow of M87*: doi.
• Sharpening up Galactic all-sky maps with complementary data - A machine learning approach: resources.

Contact information

• E-mail: web (at) ph (minus) frank (punkt) de.
• ORCID: 0000-0001-5610-3779.
• Gitlab@MPCDF: https://gitlab.mpcdf.mpg.de/pfrank.
• Google Scholar: Philipp Frank.
• LinkedIn: ph-frank.

Impressum, Datenschutzerklärung, Haftungsausschluss