Neil Kichler

M. Sc. RWTH

Contact Details

• Mail 
  Informatik 12: Software and Tools for Computational Engineering
  RWTH Aachen University
  D-52056 Aachen, Germany
• Office: Seffenter Weg 23, 52074 Aachen
• Room: 229
• Email: kichler@stce.rwth-aachen.de
• Researcher at STCE: since 2023
• Office Hours: on request
  
  

Research

• Surrogate Modelling via (second-order) Differential Machine Learning / Sobolev Training
• AD of Interval Arithmetic / McCormick Relaxations — and vice versa — on the GPU:
  
  • CuInterval: A CUDA kernel library for interval arithmetic.
  • CuMcCormick: A CUDA kernel library for McCormick relaxations.
  • CuTangent: A CUDA kernel library for subgradients using forward-mode AD.

Publications

• 2024
  • Kichler N., Afghan S., Naumann, U.: Towards Sobolev Pruning. In the Proceedings of the Platform for Advanced Scientific Computing (PASC).

• 2023
  • Atashgahi Z., Zhang X., Kichler N., Liu S., Yin L., Pechenizkiy M., Veldhuis R., Mocanu D.C.: Supervised Feature Selection with Neuron Evolution in Sparse Neural Networks. In Transactions on Machine Learning Research (TMLR).

Talks

• 2024
  • Platform for Advanced Scientific Computing (PASC) Conference, ETH Zurich, Switzerland:
    Towards Sobolev Pruning (Slides)
•  2023
  • 26th European Workshop on Automatic Differentiation, RWTH Aachen, Germany:
    Second-Order Differential Machine Learning (Website)
    
    
  • Quantminds International, London, UK:
    Second-Order Differential Machine Learning (Poster)

Teaching

• Introduction to Algorithmic Differentiation: WS24/25