arXiv:2002.01935

[refraction-ray]

• Link
• Title: Hyper-optimized tensor network contraction
• Keywords (optional):
• Authors (optional):
• Reason (optional):
• Summary (optional): very brief review and benchmarks on relevant tensor contraction path finders with some randomization added. Tested on regular graph (KaHyPar is the best in cost), planar graph and random QC. KHP seems to be consistently outperformed other scheme. Quoted as

  We find that the contractor based on hypergraph partitioning, in particular, often outperforms all other methods

• Ideas (optional):

[refraction-ray]

Relevant works on tensor contraction algorithms

in general settings, or especially in terms of quantum circuit setups

tree decompsition

Introduction to tree decom basics: tutorial

• theory related tree decomp to qc simulation with tn: https://epubs.siam.org/doi/abs/10.1137/050644756?mobileUi=0 Shi
• Quick BB: https://arxiv.org/pdf/1207.4109.pdf

TN for QC

• https://arxiv.org/pdf/1905.08394.pdf PEPS
• https://arxiv.org/pdf/2005.06787.pdf
• https://www.nature.com/articles/s41534-019-0196-1.pdf qflex and references therein
• https://arxiv.org/pdf/1905.00444 GPU version of Feynman type simulator

[refraction-ray]

Relevant packages on path finding of tensor contraction

• https://github.com/kahypar/kahypar cpp
• https://github.com/jcmgray/cotengra python interface
• https://github.com/kit-algo/flow-cutter-pace17 cpp