Pith. sign in

REVIEW 4 cited by

Tensor network ranks

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1801.02662 v2 pith:MIV2XSDK submitted 2018-01-08 math.NA cs.NA

classification math.NAcs.NA
keywords ranktensorranksmatrixthereclassicaltextscdimension
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

In problems involving approximation, completion, denoising, dimension reduction, estimation, interpolation, modeling, order reduction, regression, etc, we argue that the near-universal practice of assuming that a function, matrix, or tensor (which we will see are all the same object in this context) has \emph{low rank} may be ill-justified. There are many natural instances where the object in question has high rank with respect to the classical notions of rank: matrix rank, tensor rank, multilinear rank --- the latter two being the most straightforward generalizations of the former. To remedy this, we show that one may vastly expand these classical notions of ranks: Given any undirected graph $G$, there is a notion of $G$-rank associated with $G$, which provides us with as many different kinds of ranks as there are undirected graphs. In particular, the popular tensor network states in physics (e.g., \textsc{mps}, \textsc{ttns}, \textsc{peps}) may be regarded as functions of a specific $G$-rank for various choices of $G$. Among other things, we will see that a function, matrix, or tensor may have very high matrix, tensor, or multilinear rank and yet very low $G$-rank for some $G$. In fact the difference is in the orders of magnitudes and the gaps between $G$-ranks and these classical ranks are arbitrarily large for some important objects in computer science, mathematics, and physics. Furthermore, we show that there is a $G$ such that almost every tensor has $G$-rank exponentially lower than its rank or the dimension of its ambient space.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Two-Cut Coherence of Quintic Forms: Lifting Separations and Second-Derivative Completeness

    cs.CC 2026-08 accept novelty 7.0 of 10

    Two-cut coherence of a polynomial equals the minimal shared interface width for two degree cuts, and for quintics it is characterized up to constants by the slice rank of second derivatives.

  2. Bounding Classical and Quantum Correlations in Bayesian Networks with Quasiprobabilities

    quant-ph 2026-06 unverdicted novelty 7.0 of 10

    Quasiprobability models in Bayesian networks generalize to produce all non-signalling correlations for a broad class of networks and conjecturally recover the nested Markov model.

  3. A Provably Efficient Method for Tensor Ring Decomposition and Its Applications

    math.NA 2025-11 conditional novelty 7.0 of 10

    BLOSTR recovers exact tensor ring cores in a fixed number of algebraic steps from O(r²Σnⱼ) sampled entries, under genericity conditions and mode sizes ≥ r².

  4. Minimality of Tree Tensor Network Ranks

    math.NA 2025-09 conditional novelty 6.0 of 10

    For tree tensor networks, a bond dimension tuple is minimal if and only if at every vertex each bond dimension is no larger than the local physical dimension times the product of the other incident bond dimensions.

Pith tools