Geometry of low nonnegative rank matrix completion
Kaie Kubjas, Lilja Mets\"alampi

TL;DR
This paper investigates the geometric and combinatorial properties of low nonnegative rank matrix completion, providing characterizations and examples mainly for ranks 1 to 3, and extending geometric insights into nested polytopes.
Contribution
It offers new geometric characterizations for nonnegative rank-$r$ completions and identifies conditions under which rank and nonnegative rank completions coincide for low ranks.
Findings
Rank-1 completion characterized by rank-1 completion.
For 3x3 matrices, patterns for rank-2 and nonnegative rank-2 equivalence.
Examples showing differences between rank and nonnegative rank for rank 3.
Abstract
We study completion of partial matrices with nonnegative entries to matrices of nonnegative rank at most for some . Most of our results are for . We show that a partial matrix with nonnegative entries has a nonnegative rank-1 completion if and only if it has a rank-1 completion. This is not true in general when . For matrices, we characterize all the patterns of observed entries when having a rank-2 completion is equivalent to having a nonnegative rank-2 completion. If a partial matrix with nonnegative entries has a rank- completion that is nonnegative, where , then it has a nonnegative rank- completion. We will demonstrate examples for where this is not true. We do this by introducing a geometric characterization for nonnegative rank- completion employing families of nested polytopes which generalizes…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Optimization Algorithms Research · Sparse and Compressive Sensing Techniques · Stochastic Gradient Optimization Techniques
