Title: Orthogonal Tensor Decompositions 

Speaker:  Prof. Chao Zeng  ( University of Hong Kong )

Time: Fri, Nov.13 2020, PM: 15:00-16:00

Location:  Tencent Conference

Meeting Number: 751 219 864

Conference Link: http://meeting.tencen.com/s/MGTRM6foGkbV

Inviter: Dr. Yuanpeng Zhu

Abstract：

      The orthogonal decomposition factorizes a tensor into a sum of an orthogonal list of rank-one tensors. Unlike tensor rank, a subtensor may have a larger orthogonal tensor rank than the whole tensor, and orthogonal tensor rank is lower semicontinuous. The lower semicontinuity guarantees the existence of low orthogonal rank approximation. To fit the orthogonal decomposition, we propose ana algorithm based on the augmented Lagrangian method and guarantee the orthogonality by a novel orthogonalization procedure. Numerical eperiments show that the proposed method has great advantages over the existing methods for strongly orthogonal decompositions in terms of the approximation error.