Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a framework to bridge the "differentiability gap" in Earth system deep learning by either analytically approximating non-differentiable functions or learning differentiable surrogates using Lipschitz-constrained convolutional neural networks. They formulate an analytical approximation using temperature-controlled sigmoids and continuous logical operators to relax discrete topological operations. They demonstrate this framework by developing the Minkowski image loss, a differentiable equivalent for integral-geometric measures of surface precipitation fields, and show that while Lipschitz regularization stabilizes optimization, it can over-smooth gradient signals, limiting the recovery of localized textures.
Lipschitz-constrained neural surrogates can emulate non-differentiable scientific metrics with high accuracy, but may struggle to recover fine-grained details due to over-smoothing, highlighting a fundamental trade-off in differentiable approximation.
The ``differentiability gap'' presents a primary bottleneck in Earth system deep learning: since models cannot be trained directly on non-differentiable scientific metrics and must rely on smooth proxies (e.g., MSE), they often fail to capture high-frequency details, yielding ``blurry'' outputs. We develop a framework that bridges this gap using two different methods to deal with non-differentiable functions: the first is to analytically approximate the original non-differentiable function into a differentiable equivalent one; the second is to learn differentiable surrogates for scientific functionals. We formulate the analytical approximation by relaxing discrete topological operations using temperature-controlled sigmoids and continuous logical operators. Conversely, our neural emulator uses Lipschitz-convolutional neural networks to stabilize gradient learning via: (1) spectral normalization to bound the Lipschitz constant; and (2) hard architectural constraints enforcing geometric principles. We demonstrate this framework's utility by developing the Minkowski image loss, a differentiable equivalent for the integral-geometric measures of surface precipitation fields (area, perimeter, connected components). Validated on the EUMETNET OPERA dataset, our constrained neural surrogate achieves high emulation accuracy, completely eliminating the geometric violations observed in unconstrained baselines. However, applying these differentiable surrogates to a deterministic super-resolution task reveals a fundamental trade-off: while strict Lipschitz regularization ensures optimization stability, it inherently over-smooths gradient signals, restricting the recovery of highly localized convective textures. This work highlights the necessity of coupling such topological constraints with stochastic generative architectures to achieve full morphological realism.