Work

Advancing Tensor Network Simulations - Hardware Acceleration, Stable Optimization, and Symmetries

Year: 2024
Type: dissertation
Abstract: We discuss several algorithmic improvements for tensor network algorithms. In particular, we propose and benchmark (i) a modified truncation step that enables exploiting hardware acceleration fully, (ii) a gradient-based approach for the optimization of tensor networks for finite two-dimensional systems, and (iii) a framework to incorporate nonabelian symmetries or fermionic or anyonic exchange statistics into tensor network simulations on the tensor level.
Authors: Unfried, Jakob, Pollmann, Frank 1978-, Technische Universität München
Language: English
Open Access status: green
OpenAlex ID: W7045407873
Cites: 167
Cited by:
Related to: 9