Login

Differentiable Analysis Blueprint

Mar 5 – 6, 2026
CERN
Europe/Berlin timezone

An IRIS-HEP Blueprint Workshop

Contribution List

20 out of 20 displayed
Export to PDF
  1. 1. Welcome & Introduction to Workshop
    Lino Oscar Gerlach (Princeton University (US)), Mohamed Aly (Princeton University (US))
    3/5/26, 2:00 PM
  2. 2. Automatic Differentiation in HEP and Differentiable Systematics [30' + 15']
    Lukas Alexander Heinrich (Technische Universitat Munchen (DE))
    3/5/26, 2:10 PM
  3. 3. Differentiating Discrete Operations [10' + 5']
    Annalena Kofler (Technical University Munich)
    3/5/26, 2:55 PM

    Discrete decisions arise in simulators and analysis pipelines across disciplines such as biophysics, robotics, and HEP. Because these operations are inherently non-differentiable, the machine learning community has developed a range of methods to obtain such gradients. In this talk, I outline why a statistical perspective on gradient estimation is essential in this setting and give a brief...

    Go to contribution page
  4. 5. Discussion
    3/5/26, 3:10 PM
  5. 6. Everything that can go wrong [10'+5']
    Frederic Renner (Deutsches Elektronen-Synchrotron (DE))
    3/5/26, 4:00 PM

    My journey through pitfalls I discovered building an autodiff workflow for an ATLAS search.

    Go to contribution page
  6. 7. Adoption of Automatic Differentiation in HEP [10' + 5']
    Alexander Held (University of Wisconsin Madison (US))
    3/5/26, 4:15 PM

    How and where can AD help in HEP? What do we need to take advantage of it, how does it compare to approaches we used in previous decades? This talk will present a brief look into these questions.

    Go to contribution page
  7. 8. Discussion
    3/5/26, 4:30 PM
  8. 9. Tooling for Differentiable HEP [10' + 5']
    Lino Oscar Gerlach (Princeton University (US))
    3/5/26, 5:30 PM
  9. 23. Differentiable Binning Optimization [10' + 5']
    Nitish Kasaraguppe (RWTH Aachen (DE))
    3/5/26, 5:45 PM

    Categorizing events using discriminant observables is central to many high-energy physics analyses. Yet, bin boundaries are often chosen by hand. A simple, popular choice is to apply argmax projections of multi-class scores and equidistant binning of one-dimensional discriminants.

    This talk presents binning optimization for signal significance directly in multi-dimensional discriminants...

    Go to contribution page
  10. 10. Awkward Arrays and JAX [10' + 5']
    Ianna Osborne (Princeton University)
    3/5/26, 6:00 PM

    High-energy physics (HEP) relies on nested, variable-length (“ragged”) data structures that do not align naturally with the static, rectangular tensor abstractions assumed by most GPU compiler stacks. While Awkward Array provides a NumPy-like interface for such data, integrating it into the JAX ecosystem exposes an architectural mismatch between dynamic, offset-based structures and JAX’s...

    Go to contribution page
  11. 11. Discussion
    3/5/26, 6:15 PM
  12. 12. Closing and Summary
    Lino Oscar Gerlach (Princeton University (US)), Mohamed Aly (Princeton University (US))
    3/5/26, 7:00 PM
  13. 13. Introduction to the day
    Lino Oscar Gerlach (Princeton University (US)), Mohamed Aly (Princeton University (US))
    3/6/26, 2:00 PM
  14. 14. Differentiable Statistical Ecosystem with JAX [10' + 5']
    Peter Fackeldey (Princeton University (US))
    3/6/26, 2:10 PM

    The propagation of gradients with backpropagation through a HEP analysis for end-to-end optimization begins at the last step of a physics analysis: the statistical measurement. Therefore, it is crucial to have statistical tools that are fully differentiable in order to calculate gradients with respect to the final physical measurement. This contribution provides an overview of how such fully...

    Go to contribution page
  15. 15. Differentiable Programming in C++ and ROOT with Clad [25' + 5']
    Jonas Rembser (CERN), Vassil Vasilev (Princeton University (US))
    3/6/26, 2:25 PM

    In this presentation, we make the case for differential programming in C++ for High Energy Physics. We will first introduce source code-transformation-based Automatic Differentiation (AD) with Clad, a Clang compiler plugin. Some success stories of how this is used for statistical analysis in ROOT are presented, including using Clad to differentiate through statistical likelihoods in RooFit and...

    Go to contribution page
  16. 18. Discussion
    3/6/26, 2:55 PM
  17. 17. Differentiable Detector Simulation with GEANT [15' + 5']
    Jeffrey Krupa (SLAC)
    3/6/26, 3:40 PM

    Applying automatic differentiation (AD) to particle simulations such as Geant4 opens the possibility of gradient-based optimization for detector design and parameter tuning in high-energy physics. We extend our previous work on differentiable Geant simulations by incorporating multiple Coulomb scattering into the physics model, moving closer to realistic detector modeling. The inclusion of...

    Go to contribution page
  18. 19. End-to-End Differentiable Analysis in IceCube [20' + 10']
    Oliver Janik (FAU Erlangen-Nürnberg (DE))
    3/6/26, 4:00 PM

    Measurements of the astrophysical neutrino flux with the IceCube Neutrino Observatory traditionally rely on binned forward-folding likelihood analyses. These methods require Monte Carlo simulations to predict event distributions. Limited Monte Carlo statistics restrict the dimensionality of the binning and therefore the amount of exploitable information.
    This talk presents a fully...

    Go to contribution page
  19. 21. Automatic Differentiation Beyond HEP [20' + 10']
    Giacomo Acciarini (European Space Agency (ESA))
    3/6/26, 4:30 PM

    Differentiable programming is advancing scientific computing by enabling gradients to flow through complex numerical models. In spaceflight mechanics, a field governed by nonlinear dynamics, uncertainty, and strict operational constraints, this approach opens new avenues for optimization, state estimation, uncertainty quantification, and decision-making.
    In this talk, I will present our...

    Go to contribution page
  20. 22. Closing, Call for Action and Future Steps
    Lino Oscar Gerlach (Princeton University (US)), Mohamed Aly (Princeton University (US))
    3/6/26, 5:45 PM