The Dark Side of the Forces: Assessing Non-Conservative Force Models for Atomistic Machine Learning
Summary of Bigi et al.'s 2025 paper that critically evaluates the trend of using non-conservative force models in atomistic machine learning, …
Icml-2025
4 posts tagged with icml-2025.
Efficient and Scalable Density Functional Theory Hamiltonian Prediction through Adaptive Sparsity
Summary of Luo et al.'s 2025 paper introducing SPHNet, an SE(3)-equivariant network that uses adaptive sparsity to dramatically improve the efficiency …
Learning Smooth and Expressive Interatomic Potentials for Physical Property Prediction
Summary of Fu et al.'s 2025 paper that bridges the gap between test-set accuracy and downstream task performance for MLIPs by proposing energy …
Beyond Atoms: Enhancing Molecular Pretrained Representations with 3D Space Modeling
Summary of Lu et al.'s 2025 paper introducing SpaceFormer, a Transformer-based framework that models the entire 3D space around a molecule—not just …