On Multilinear Forms: Bias, Correlation, and Tensor Rank

@inproceedings{Bhrushundi2018OnMF,
  title={On Multilinear Forms: Bias, Correlation, and Tensor Rank},
  author={Abhishek Bhrushundi and Prahladh Harsha and Pooya Hatami and Swastik Kopparty and Mrinal Kumar},
  booktitle={Electron. Colloquium Comput. Complex.},
  year={2018}
}
In this paper, we prove new relations between the bias of multilinear forms, the correlation between multilinear forms and lower degree polynomials, and the rank of tensors over $GF(2)= \{0,1\}$. We show the following results for multilinear forms and tensors. 1. Correlation bounds : We show that a random $d$-linear form has exponentially low correlation with low-degree polynomials. More precisely, for $d \ll 2^{o(k)}$, we show that a random $d$-linear form $f(X_1,X_2, \dots, X_d) : \left(GF(2… 
6 Citations
The analytic rank of tensors and its applications
The analytic rank of tensors and its applications, Discrete Analysis 2019:7, 10 pp. There are several arguments in additive combinatorics concerning two-variable functions taking values in a field
Geometric rank of tensors and subrank of matrix multiplication
TLDR
It is proved that the geometric rank is an upper bound on the subrank of tensors and the independence number of hypergraphs, and it is related to the analytic rank of Gowers and Wolf in an asymptotic fashion.
Subrank and Optimal Reduction of Scalar Multiplications to Generic Tensors
TLDR
This paper proves for a generic bilinear map T : V × V → V where dim( V ) = n that θ ( √ n ) independent scalar multiplications can be reduced to T .
C O ] 1 4 O ct 2 01 8 The analytic rank of tensors and its applications
The analytic rank of a tensor, first defined by Gowers and Wolf in the context of higher-order Fourier analysis, is defined to be the logarithm of the bias of the tensor. We prove that it is a
Subspaces of tensors with high analytic rank
It is shown that for any subspace V⊆Fn×⋯×np of d-tensors, if dim(V)≥tnd−1, then there is subspace W⊆V of dimension at least t/(dr)−1 whose nonzero elements all have analytic rank Ωd,p(r). As an
SUBSPACES OF TENSORS WITH HIGH ANALYTIC RANK
It is shown that for any subspace V ⊆ Fn×···×n p of d-tensors, if dim(V ) ≥ tnd−1, then there is subspace W ⊆ V of dimension t/(dr) − 1 whose nonzero elements all have analytic rank Ωd,p(r). As an

References

SHOWING 1-10 OF 26 REFERENCES
Tensor Rank: Some Lower and Upper Bounds
TLDR
This work explores tensor rank lower and upper bounds, focusing on explicit tensors, and shows that many permutation tensors have far from maximal rank, which is very different from the matrix case and thus eliminates many natural candidates for high Tensor rank.
The Chasm at Depth Four, and Tensor Rank : Old results, new insights
TLDR
This paper abstracts the main ingredient to apply it to formulas and constant depth circuits, and shows more structured depth reductions for them, and extends this result for homogeneous formulas to show that the connection holds for any $d$ such that $\omega(1) d \leq n^{o(1)}$.
The analytic rank of tensors is subadditive, and its applications
TLDR
It is proved that the analytic rank of a tensor lower bounds the slice rank and partition rank of an tensor, quantities that arise in the study of Ramsey problems in product spaces and suggests a new approach to study some open problems in this domain.
Random low-degree polynomials are hard to approximate
TLDR
It is proved that almost all degree d polynomials have only an exponentially small correlation with all polynmials of degree at most d − 1, for all degrees d up to Θ(n).
Lower bounds for matrix product
  • Amir Shpilka
  • Computer Science, Mathematics
    Proceedings 2001 IEEE International Conference on Cluster Computing
  • 2001
We prove lower bounds on the number of product gates in bilinear and quadratic circuits that compute the product of two n /spl times/ n matrices over finite fields. In particular we obtain the
A Lower Bound on the Complexity of Polynomial Multiplication over Finite Fields
  • M. Kaminski
  • Mathematics, Computer Science
    SIAM J. Comput.
  • 2005
It is shown that computing the coefficients of the product of two degree-n polynomials over a q-element field by means of a quadratic algorithm requires at least $(3+
Affine dispersers from subspace polynomials
TLDR
The main novelty in the construction lies in the method of proof, which relies on elementary properties of <i>subspace polynomials</i>.
Barriers for Rank Methods in Arithmetic Complexity
TLDR
This paper addresses rank methods, which were long recognized as encompassing and abstracting almost all known arithmetic lower bounds to-date, including the most recent impressive successes, and results are barriers to these methods.
Norms, XOR Lemmas, and Lower Bounds for Polynomials and Protocols
TLDR
It is shown that if a Boolean function has correlation at most e 1/2 with either of these models, then the correlation of the parity of its values on m independent instances drops exponentially with m, and for polynomials over GF(2) of degree d, the correlation drops to exp m/4 d.
Guest Column: correlation bounds for polynomials over {0 1}
This article is a unified treatment of the state-of-the-art on the fundamental challenge of exhibiting explicit functions that have small correlation with low-degree polynomials over {0 1 }. It
...
...