We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: TenIPS: Inverse Propensity Sampling for Tensor Completion

Abstract: Tensors are widely used to represent multiway arrays of data. The recovery of missing entries in a tensor has been extensively studied, generally under the assumption that entries are missing completely at random (MCAR). However, in most practical settings, observations are missing not at random (MNAR): the probability that a given entry is observed (also called the propensity) may depend on other entries in the tensor or even on the value of the missing entry. In this paper, we study the problem of completing a partially observed tensor with MNAR observations, without prior information about the propensities. To complete the tensor, we assume that both the original tensor and the tensor of propensities have low multilinear rank. The algorithm first estimates the propensities using a convex relaxation and then predicts missing values using a higher-order SVD approach, reweighting the observed tensor by the inverse propensities. We provide finite-sample error bounds on the resulting complete tensor. Numerical experiments demonstrate the effectiveness of our approach.
Comments: AISTATS 2021
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC)
Cite as: arXiv:2101.00323 [stat.ML]
  (or arXiv:2101.00323v3 [stat.ML] for this version)

Submission history

From: Chengrun Yang [view email]
[v1] Fri, 1 Jan 2021 22:13:19 GMT (3861kb,D)
[v2] Mon, 1 Mar 2021 23:34:26 GMT (3985kb,D)
[v3] Thu, 22 Apr 2021 15:41:38 GMT (3985kb,D)

Link back to: arXiv, form interface, contact.