Tag
This paper introduces a mesh-free multiresolution deep energy method using neural networks for phase-field modeling of brittle fracture, demonstrating performance comparable to finite element methods in crack propagation simulations.
This paper introduces a B-spline-based decoupling framework for compressing transformer models, with a robust alternating least-squares algorithm (R-CMTF-BSD) that achieves substantial parameter reduction while maintaining competitive accuracy on Vision and Swin Transformer architectures.