Publications 2024

• TrAct: Making First-layer Pre-Activations Trainable
  Felix Petersen, Christian Borgelt, and Stefano Ermon
  Proc. Int. Conf. on Neural Information Processing 37 (NeurIPS 2024, Vancouver, Canada)
  doi:10.48550/arXiv.2410.23970
  neurips_24c.pdf (2345 kb) (10/17 pages)
• Newton Losses: Using Curvature Information for Learning with Differentiable Algorithms
  Felix Petersen, Christian Borgelt, Tobias Sutter, Hilde Kuehne, Oliver Deussen, and Stefano Ermon
  Proc. Int. Conf. on Neural Information Processing 37 (NeurIPS 2024, Vancouver, Canada)
  doi:10.48550/arXiv.2410.19055
  neurips_24b.pdf (763 kb) (10/21 pages
• Convolutional Differentiable Logic Gate Networks
  Felix Petersen, Hilde Kuehne, Christian Borgelt, Julian Welzel, and Stefano Ermon
  Proc. Int. Conf. on Neural Information Processing 37 (NeurIPS 2024, Vancouver, Canada)
  doi:10.48550/arXiv.2411.04732
  neurips_24a.pdf (1315 kb) (10/17 pages)
• Uncertainty Quantification via Stable Distribution Propagation
  Felix Petersen, Aashwin Ananda Mishra, Hilde Kuehne, Christian Borgelt, Oliver Deussen, and Mikhail Yurochkin
  Proc. 12th Int. Conf. on Learning Representations (ICLR 2024, Vienna, Austria)
  doi:10.48550/arXiv.2402.08324 openreview.net
  iclr_24.pdf (1319 kb) (9/27 pages)

The publications made available on these pages as PDF or Postcript files are, unless published as Open Access, preliminary draft versions that may deviate from the finally published printed versions that are subject to copyright restrictions.