Search papers, labs, and topics across Lattice.
University of Electronic Science and Technology of China
4
0
3
Automating rank determination in tensor-based neural networks could revolutionize their application in complex physics simulations.
A new tensor completion method achieves superior performance by leveraging a flexible nonconvex surrogate that accurately captures tensor structure.
Stop being limited by fixed tensor decomposition families: TenExp dynamically mixes and matches decompositions to better capture low-rank data structures.
Unlock the potential of continuous tensor representations with neural operators, achieving more faithful representations of complex real-world data and outperforming classic discrete and continuous tensor methods.
