PH

Philipp Hähnel

PhD, Theoretical Physicist and AI Researcher

about

Philipp Hähnel

Name: Philipp Hähnel

Nationality: German

About Me

I am serious. Humorous, (German-style) funny, precise, friendly, benevolent, forgiving, sharp, focused, efficient, optimized, interested, creative, smart and independent.


Research Interests

artificial general intelligence | reinforcement learning | deep learning | symmetries and invariance | multi-agent systems | adversarial learning | mathematical foundations

computational genomics | positive selection

higher spin theories | gauge/gravity duality and holography | quantum gravity | twistor theory | gauge field theory


Hobbies

dancing | story writing | world building | role playing | hiking | piano


Since long ago mankind has searched and found inspiration in the new and unexplainable. This instinct has led us to develop many a model, describing experimental data and foremost predicting the outcome of new measurements, thus increasing our understanding of the numerous mysteries of nature. Layer by layer, each mathematical abstraction in our models endowed a simplification in its description which yet became harder and harder to grasp. Today, the search for a model of everything leads us scientists only deeper into the maze of complex theories, in which we are still zealously fighting the minotaur who is guarding its secret. For we all have the hope that one day, after the smoke of uncountably many notational conventions has cleared and the dust of (un)pleasingly bloody calculations has settled, the vision becomes clear, revealing the very relations that we were looking for so long.

education

Skills

Professional

Python
2 years
Mathematica
13 years
LaTeX
14 years
screwing around with
other people's code
forever
Blues Dancing
many years

Intermediate

Tensorflow
1 year
MongoDB
>1 year
Matlab / Scilab
3 years
Photoshop
3 years
Piano
decades

Beginner

Fiddle
3 years (a few years ago)

Education

2014-2018

Doctor of Philosophy, Mathematics

Trinity College Dublin, Ireland

Higher Spin Theories in Twistor Space

Using the twistor formalism, I was able to formulate an action principle for conformal higher spin theory. The action poses as generating functional to compute scattering amplitudes for this theory, which had been and still is a difficult problem since its conception.

2011-2014

Master of Science, Physics

Humboldt University of Berlin, Germany

The One-Loop Effective Action of N=4 SYM-Type Theories

I developed a Mathematica program to automate the computation of the first-order renormalization constants of many Lagrangian theories. The program computes the divergent part of the first-order quantum corrections to the effective action using the background field method and dimensional regularization.

2007-2011

Bachelor of Science, Physics

Humboldt University of Berlin, Germany

Minimal Surfaces in Anti-de Sitter Spaces

I studied conformal transformations of minimal surfaces in AdS space that correspond to gluon scattering amplitudes via the AdS/CFT correspondence.

patents

Patent Applications

  1. Deep learning for PDE-based Models

    P. Haehnel, Jakub Marecek, Julien Monteil, Fearghal O'Donncha: patent application on Deep Learning for PDE-based Models, serial no. 16/121,315, filed through I.B.M. at the USPTO

publications

Publications

  1. Scaling up Deep Learning for PDE-based Models

    P. Haehnel, Jakub Marecek, Julien Monteil, and Fearghal O'Donncha: arXiv:1810.09425 [cs.LG],
  2. Higher Spin Theories in Twistor Space

    P. Haehnel: TCD handle, Trinity College Dublin.School of Mathematics.MATHEMATICS, 2018
  3. Conformal higher spin scattering amplitudes from twistor space

    T. Adamo, P. Haehnel, and T. McLoughlin: arXiv:1611.06200 [hep-th], JHEP 1704:021, 2017
  4. Conformal higher spin theory and twistor space actions

    P. Haehnel, T. McLoughlin: arXiv:1604.08209 [hep-th], J. Phys. A: Math. Theor. 50 485401, (Selected for Journal of Physics A Highlights of 2017)
  5. On force-field models of the spacecraft flyby anomaly

    W. Hasse, E. Birsin, and P. Haehnel: arXiv:0903.0109 [gr-qc]

experience

Work Experience

2019 - present

Research Fellow

Dana-Farber Cancer Institute, Harvard T.H. Chan School of Public Health

Developing tools for studying differential positive selection in cancer, using Bayesian inference and generative modeling

2018

Fellow

Institute for Pure and Applied Mathematics, UCLA

Science at Extreme Scales: Where Big Data Meets Large-Scale Computing

Leading a study group focusing on exploring and exploiting symmetries and invariants in algorithm design and data analysis; partaking in study groups on scalability, model development, and reinforcement learning

2018

Research Scientist Intern

I.B.M., Dublin, Ireland

Deep learning for pollution modelling and forecasting: using traffic and weather data, to replace an expensive PDE-based model with a scalable ML model by imposing consistency constraints across the boundaries of a mesh that decomposes a large domain

2014-2018

Teaching assistant

Trinity College Dublin, Ireland

Differential Geometry, General Relativity, Quantum Mechanics, Advanced Calculus, Classical Field Theory, Classical Electrodynamics

2011-2013

Teaching assistant

Humboldt University of Berlin, Germany

Quantum Field Theory I & II, Linear Algebra and Analytical Geometry I & II

2006-2012

Organization of seminars for secondary school students

Astrometrie, Wilhelm Foerster Observatory, Berlin;
seminars were held at the Technical University of Berlin

Introduction to General Relativity, Physics of the Sun, The EPR-Paradox, Anomalies in the Solar System, Gravitational Lenses, Physics of Stars, Recent Cosmology, Introduction to Quantum Physics, Black Holes

2010-2011

Student Assistant

Neurorobotics Research Laboratory, Department of Informatics, HU Berlin

Software engineering: 2D physics simulator for exploration of autonomous robot designs

2007

Student Internship

Fraunhofer Institute for Open Communication Systems, Berlin

Data analysis: classifying neural signal data using k-fold cross-validation

2005

Student Internship

German Aerospace Center (DLR), Berlin

Image analysis: calculating atmospheric height of dust clouds on Mars

Side Projects

Honors and Awards

  •   Over 10 invited talks  
    and posters

    2014-present
  •   'String Theory Universe'  
    travel grant

    2015
  • Scholarship of the
      German National Academic Foundation  

    2007-2011
  •   School's best graduation  
    in physics  

    2007
  • 3rd place in national round
      of the 42nd competition 'Jugend forscht'  

    2007
  • 2nd place in Berlin round
      of the 42nd competition 'Jugend forscht'  

    2007

Silliness

  • xkcd

  • SMBC

  • Sequential Art

  • Stand Still Stay Silent

  • Adventures of God

  • Tower of God

contact

Let's Talk

Location: Boston
Email: phahnel [at] hsph.harvard.edu
Last Update: July 10, 2019