US 12,019,959 B2
Distributed tensor network contraction scheme with splitting based on dynamic ordering
Jiachen Huang, San Mateo, CA (US); Fang Zhang, San Mateo, CA (US); and Jianxin Chen, Kirkland, WA (US)
Assigned to Alibaba Group Holding Limited, Grand Cayman (KY)
Filed by ALIBABA GROUP HOLDING LIMITED, Grand Cayman (KY)
Filed on Dec. 31, 2020, as Appl. No. 17/139,281.
Claims priority of provisional application 62/957,442, filed on Jan. 6, 2020.
Prior Publication US 2021/0209270 A1, Jul. 8, 2021
Int. Cl. G06F 30/20 (2020.01); G06F 17/16 (2006.01); G06N 10/00 (2022.01)
CPC G06F 30/20 (2020.01) [G06F 17/16 (2013.01); G06N 10/00 (2019.01)] 17 Claims
OG exemplary drawing
 
1. A method for performing contraction on a tensor network, comprising:
receiving, by a system, a tensor network comprising a plurality of tensors and a plurality of edges among the plurality of tensors, wherein the edges comprise one or more open edges and one or more closed edges, and each edge is associated with a plurality of index elements;
determining a contraction order of the tensor network by:
creating a virtual tensor for connecting the one or more open edges;
generating an intermediate tensor network, the intermediate tensor network having the one or more open edges closed by the virtual tensor;
performing tree decomposition on the intermediate tensor network to construct a tree; and
determining the contraction order of the tensor network based on the tree;
determining, among the plurality of edges, one or more edges for generating a plurality of sub-networks based on the tensor network; and
distributing the plurality of sub-networks to a plurality of computing nodes of the system to perform, by the plurality of computing nodes, contraction on the plurality of sub-networks based on the contraction order.