2 papers
cs.LG2026
Drifting Fields are not Conservative
Leonard T. Franz, Sebastian Hoffmann, Tim Weiland +2
Drifting models have recently gained attention for generating high-quality samples in a single forward pass. During training, they learn a push-forward map by following a vector-va…
cs.LG2026
SoftJAX & SoftTorch: Empowering Automatic Differentiation Libraries with Informative Gradients
Anselm Paulus, A. René Geist, VÃt Musil +3
Automatic differentiation (AD) frameworks such as JAX and PyTorch have enabled gradient-based optimization for a wide range of scientific fields. Yet, many "hard" primitives in the…