Bias implies low rank for quartic polynomials
Amichai Lampert

TL;DR
This paper explores the structure of degree four polynomials over finite fields, showing that high bias implies a low-rank representation with a logarithmic bound on the number of lower-degree polynomials involved.
Contribution
It advances understanding by establishing a log-polynomial bound on the number of lower-degree polynomials needed to represent biased quartic polynomials, improving previous polynomial bounds.
Findings
High bias implies a simple low-rank structure for quartic polynomials.
The number of lower degree polynomials needed is at most log-polynomial in 1/δ.
The result aligns with recent independent proofs for arbitrary degree polynomials.
Abstract
We investigate the structure of polynomials of degree four in many variables over a fixed prime field . In 2007, Green and Tao proved that if a polynomial is poorly distributed, then it is a function of a few polynomials of smaller degree. In 2009, Haramaty and Shpilka found an effective bound for of degree four: If , then the number of lower degree polynomials required is at most polynomial in and has a simple presentation as a sum of their products. We make a step towards showing that in fact the number of lower degree polynomials required is at most log-polynomial in , with the same simple presentation of . This result was a Master's thesis supervised by T. Ziegler at the Hebrew University of Jerusalem, submitted in October 2018. A log-polynomial bound for…
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
TopicsCoding theory and cryptography · Finite Group Theory Research · Limits and Structures in Graph Theory
