The Complexity of Contracting Planar Tensor Network

01/28/2020
by   Liu Ying, et al.
0

Tensor networks have been an important concept and technique in many research areas such as quantum computation and machine learning. We study the complexity of evaluating the value of a tensor network. This is also called contracting the tensor network. In this article, we focus on computing the value of a planar tensor network where every tensor specified at a vertex is a Boolean symmetric function. We design two planar gadgets to obtain a sub-exponential time algorithm. The key is to remove high degree vertices while essentially not changing the size of the tensor network. The algorithm runs in time (O(√(|V|))). Furthermore, we use a counting version of the Sparsification Lemma to prove a matching lower bound (Ω(√(|V|))) assuming #ETH holds.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
12/27/2017

Tensor network complexity of multilinear maps

We study tensor networks as a model of arithmetic computation for evalua...
research
02/20/2019

Finding big matchings in planar graphs quickly

It is well-known that every n-vertex planar graph with minimum degree 3 ...
research
04/12/2023

A Hall-type theorem with algorithmic consequences in planar graphs

Given a graph G=(V,E), for a vertex set S⊆ V, let N(S) denote the set of...
research
11/29/2019

Tight Bounds for Planar Strongly Connected Steiner Subgraph with Fixed Number of Terminals (and Extensions)

(see paper for full abstract) Given a vertex-weighted directed graph G...
research
07/06/2022

Tensor networks in machine learning

A tensor network is a type of decomposition used to express and approxim...
research
02/06/2022

The Exponential-Time Complexity of the complex weighted #CSP

In this paper, I consider a fine-grained dichotomy of Boolean counting c...
research
02/27/2017

Faster Tensor Canonicalization

The Butler-Portugal algorithm for obtaining the canonical form of a tens...

Please sign up or login with your details

Forgot password? Click here to reset