Timezone: »

 
Poster
Interpolating Convex and Non-Convex Tensor Decompositions via the Subspace Norm
Qinqing Zheng · Ryota Tomioka

Tue Dec 08 04:00 PM -- 08:59 PM (PST) @ 210 C #79 #None
in Poster Session »
We consider the problem of recovering a low-rank tensor from its noisy observation. Previous work has shown a recovery guarantee with signal to noise ratio $O(n^{\ceil{K/2}/2})$ for recovering a $K$th order rank one tensor of size $n\times \cdots \times n$ by recursive unfolding. In this paper, we first improve this bound to $O(n^{K/4})$ by a much simpler approach, but with a more careful analysis. Then we propose a new norm called the \textit{subspace} norm, which is based on the Kronecker products of factors obtained by the proposed simple estimator. The imposed Kronecker structure allows us to show a nearly ideal $O(\sqrt{n}+\sqrt{H^{K-1}})$ bound, in which the parameter $H$ controls the blend from the non-convex estimator to mode-wise nuclear norm minimization. Furthermore, we empirically demonstrate that the subspace norm achieves the nearly ideal denoising performance even with $H=O(1)$.
[ PDF] 

Author Information

Qinqing Zheng (University of Chicago)
Ryota Tomioka (Microsoft Research Cambridge)

More from the Same Authors