Tensor Ring Decomposition: Optimization Landscape and One-loop Convergence of Alternating Least Squares

Abstract

In this work, we study the tensor ring decomposition and its associated numerical algorithms. We establish a sharp transition of algorithmic difficulty of the optimization problem as the bond dimension increases: On one hand, we show the existence of spurious local minima for the optimization landscape even when the tensor ring format is much over-parameterized, i.e., with bond dimension much larger than that of the true target tensor. On the other hand, when the bond dimension is further increased, we establish one-loop convergence for alternating least square algorithm for tensor ring decomposition. The theoretical results are complemented by numerical experiments for both local minimum and one-loop convergence for the alternating least square algorithm.

References in corpus (13)

• The density-matrix renormalization group
• Entanglement renormalization
• An Area Law for One Dimensional Quantum Systems
• DMRG and periodic boundary conditions: a quantum information perspective
• Renormalization algorithms for Quantum-Many Body Systems in two and higher dimensions
• The density matrix renormalization group for ab initio quantum chemistry
• Tensor Ring Decomposition
• No Spurious Local Minima in Nonconvex Low Rank Problems: A Unified Geometric Analysis
• Exponential Decay of Correlations Implies Area Law
• Spectral tensor-train decomposition
• On the Convergence of Alternating Least Squares Optimisation in Tensor Format Representations
• Tensor network ranks
• The trouble with tensor ring decompositions