Searcharxiv⌕ Search

arXiv subjects

J. Maurice Rojas

Publications and source records attributed to J. Maurice Rojas.

At least 19 recordsLinked to original sources
arXiv2025

Viro's patchworking and the signed reduced A-discriminant

Computing the isotopy type of a hypersurface, defined as the positive real zero set of a multivariate polynomial, is a challenging problem in real algebraic geometry. We focus on the case where the defining polynomial has combinatorially restricted exponent vectors and fixed coefficient signs, enabling faster computation of the isotopy type. In particular, Viro's patchworking provides a polyhedral complex that has the same isotopy type as the hypersurface, for certain choices of the coefficients. So we present properties of the signed support, focussing mainly on the case of n-variate (n+3)-nomials, that ensure all possible isotopy types can be obtained via patchworking. To prove this, we study the signed reduced A-discriminant and show that it has a simple structure if the signed support satisfies some combinatorial conditions.

math.AG↗
arXiv2025

Trinomials and Deterministic Complexity Limits for Real Solving

We detail an algorithm that -- for all but a $\frac{1}{Ω(\log(dH))}$ fraction of $f\in\mathbb{Z}[x]$ with exactly $3$ monomial terms, degree $d$, and all coefficients in $\{-H,\ldots, H\}$ -- produces an approximate root (in the sense of Smale) for each real root of $f$ in deterministic time $\log^{4+o(1)}(dH)$ in the classical Turing model. (Each approximate root is a rational with logarithmic height $O(\log(dH))$.) The best previous deterministic bit complexity bounds were exponential in $\log d$. We then relate this to Koiran's Trinomial Sign Problem (2017): Decide the sign of a degree $d$ trinomial $f\in\mathbb{Z}[x]$ with coefficients in $\{-H,\ldots,H\}$, at a point $r\!\in\!\mathbb{Q}$ of logarithmic height $\log H$, in (deterministic) time $\log^{O(1)}(dH)$. We show that Koiran's Trinomial Sign Problem admits a positive solution, at least for a fraction $1-\frac{1}{Ω(\log(dH))}$ of the inputs $(f,r)$.

math.AG↗
arXiv2025

Optimal Bounds for the Number of Pieces of Near-Circuit Hypersurfaces

Suppose $f$ is a polynomial in $n$ variables with real coefficients, exactly $n+k$ monomial terms, and Newton polytope of positive volume. Estimating the number of connected components of the positive zero set of $f$ is a fundamental problem in real algebraic geometry, with applications in computational complexity and topology. We prove that the number of connected components is at most $3$ when $k\!=\!3$, settling an open question from Fewnomial Theory. Our results also extend to exponential sums with real exponents. A key contribution here is a deeper analysis of the underlying $\mathcal{A}$-discriminant curves, which should be of use for other quantitative geometric problems.

math.AG↗
arXiv2024

On the BCSS Proof of the Fundamental Theorem of Algebra

Section 10.4 of the 1998 Springer-Verlag book {\em Complexity and Real Computation}, by Blum, Cucker, Shub, and Smale, contains a particularly elegant proof of the Fundamental Theorem of Algebra: The central idea of the proof naturally leads to a homotopy continuation algorithm for finding the roots of univariate polynomials, and extends naturally to a proof of Bézout's Theorem (on the number of roots of systems of $n$ equations in $n$ unknowns). We present a more detailed version of the BCSS Proof which is hopefully useful for students and researchers not familiar with algebraic geometry. So while there are no new results in this paper, the exposition is arguably more elementary. Any errors here are solely the responsibility of the current author.

math.AG↗
arXiv2022

Sturm's Theorem with Endpoints

Sturm's Theorem is a fundamental 19th century result relating the number of real roots of a polynomial $f$ in an interval to the number of sign alternations in a sequence of polynomial division-like calculations. We provide a short direct proof of Sturm's Theorem, including the numerically vexing case (ignored in many published accounts) where an interval endpoint is a root of $f$.

cs.SC↗
arXiv2021

Root Repulsion and Faster Solving for Very Sparse Polynomials Over $p$-adic Fields

For any fixed field $K\!\in\!\{\mathbb{Q}_2,\mathbb{Q}_3,\mathbb{Q}_5, \ldots\}$, we prove that all polynomials $f\!\in\!\mathbb{Z}[x]$ with exactly $3$ (resp. $2$) monomial terms, degree $d$, and all coefficients having absolute value at most $H$, can be solved over $K$ within deterministic time $\log^{7+o(1)}(dH)$ (resp. $\log^{2+o(1)}(dH)$) in the classical Turing model: Our underlying algorithm correctly counts the number of roots of $f$ in $K$, and for each such root generates an approximation in $\mathbb{Q}$ with logarithmic height $O(\log^3(dH))$ that converges at a rate of $O\!\left((1/p)^{2^i}\right)$ after $i$ steps of Newton iteration. We also prove significant speed-ups in certain settings, a minimal spacing bound of $p^{-O(p\log^2_p(dH)\log d)}$ for distinct roots in $\mathbb{C}_p$, and even stronger repulsion when there are nonzero degenerate roots in $\mathbb{C}_p$: $p$-adic distance $p^{-O(\log_p(dH))}$. On the other hand, we prove that there is an explicit family of tetranomials with distinct nonzero roots in $\mathbb{Z}_p$ indistinguishable in their first $Ω(d\log_p H)$ most significant base-$p$ digits.

math.NT↗
arXiv2021

Counting Real Roots in Polynomial-Time for Systems Supported on Circuits

Suppose $A=\{a_1,\ldots,a_{n+2}\}\subset\mathbb{Z}^n$ has cardinality $n+2$, with all the coordinates of the $a_j$ having absolute value at most $d$, and the $a_j$ do not all lie in the same affine hyperplane. Suppose $F=(f_1,\ldots,f_n)$ is an $n\times n$ polynomial system with generic integer coefficients at most $H$ in absolute value, and $A$ the union of the sets of exponent vectors of the $f_i$. We give the first algorithm that, for any fixed $n$, counts exactly the number of real roots of $F$ in in time polynomial in $\log(dH)$.

math.AG↗
arXiv2021

A complexity chasm for solving univariate sparse polynomial equations over $p$-adic fields

We reveal a complexity chasm, separating the trinomial and tetranomial cases, for solving univariate sparse polynomial equations over certain local fields. First, for any fixed field $K\in\{\mathbb{Q}_2,\mathbb{Q}_3,\mathbb{Q}_5,\ldots\}$, we prove that any polynomial $f\in\mathbb{Z}[x]$ with exactly $3$ monomial terms, degree $d$, and all coefficients having absolute value at most $H$, can be solved over $K$ in deterministic time $O(\log^{O(1)}(dH))$ in the classical Turing model. (The best previous algorithms were of complexity exponential in $\log d$, even for just counting roots in $\mathbb{Q}_p$.) In particular, our algorithm generates approximations in $\mathbb{Q}$ with bit-length $O(\log^{O(1)}(dH))$ to all the roots of $f$ in $K$, and these approximations converge quadratically under Newton iteration. On the other hand, we give a unified family of tetranomials requiring $Ω(d\log H)$ digits to distinguish the base-$p$ expansions of their roots in $K$.

math.NT↗
arXiv2021

Tropical Varieties for Exponential Sums

We study the complexity of approximating complex zero sets of certain $n$-variate exponential sums. We show that the real part, $R$, of such a zero set can be approximated by the $(n-1)$-dimensional skeleton, $T$, of a polyhedral subdivision of $\mathbb{R}^n$. In particular, we give an explicit upper bound on the Hausdorff distance: $Δ(R,T) =O\left(t^{3.5}/δ\right)$, where $t$ and $δ$ are respectively the number of terms and the minimal spacing of the frequencies of $g$. On the side of computational complexity, we show that even the $n=2$ case of the membership problem for $R$ is undecidable in the Blum-Shub-Smale model over $\mathbb{R}$, whereas membership and distance queries for our polyhedral approximation $T$ can be decided in polynomial-time for any fixed $n$.

math.AG↗
arXiv2021

Sub-Linear Point Counting for Variable Separated Curves over Prime Power Rings

Let $k,p\in \mathbb{N}$ with $p$ prime and let $f\in\mathbb{Z}[x_1,x_2]$ be a bivariate polynomial with degree $d$ and all coefficients of absolute value at most $p^k$. Suppose also that $f$ is variable separated, i.e., $f=g_1+g_2$ for $g_i\in\mathbb{Z}[x_i]$. We give the first algorithm, with complexity sub-linear in $p$, to count the number of roots of $f$ over $\mathbb{Z}$ mod $p^k$ for arbitrary $k$: Our Las Vegas randomized algorithm works in time $(dk\log p)^{O(1)}\sqrt{p}$, and admits a quantum version for smooth curves working in time $(d\log p)^{O(1)}k$. Save for some subtleties concerning non-isolated singularities, our techniques generalize to counting roots of polynomials in $\mathbb{Z}[x_1,\ldots,x_n]$ over $\mathbb{Z}$ mod $p^k$. Our techniques are a first step toward efficient point counting for varieties over Galois rings (which is relevant to error correcting codes over higher-dimensional varieties), and also imply new speed-ups for computing Igusa zeta functions of curves. The latter zeta functions are fundamental in arithmetic geometry.

math.NT↗
arXiv2020

Computing zeta functions of large polynomial systems over finite fields

In this paper, we improve the algorithms of Lauder-Wan \cite{LW} and Harvey \cite{Ha} to compute the zeta function of a system of $m$ polynomial equations in $n$ variables over the finite field $\FF_q$ of $q$ elements, for $m$ large. The dependence on $m$ in the original algorithms was exponential in $m$. Our main result is a reduction of the exponential dependence on $m$ to a polynomial dependence on $m$. As an application, we speed up a doubly exponential time algorithm from a software verification paper \cite{BJK} (on universal equivalence of programs over finite fields) to singly exponential time. One key new ingredient is an effective version of the classical Kronecker theorem which (set-theoretically) reduces the number of defining equations for a "large" polynomial system over $\FF_q$ when $q$ is suitably large.

math.NT↗
arXiv2019

Randomized Polynomial-Time Root Counting in Prime Power Rings

Suppose $k,p\!\in\!\mathbb{N}$ with $p$ prime and $f\!\in\!\mathbb{Z}[x]$ is a univariate polynomial with degree $d$ and all coefficients having absolute value less than $p^k$. We give a Las Vegas randomized algorithm that computes the number of roots of $f$ in $\mathbb{Z}/\!\left(p^k\right)$ within time $d^3(k\log p)^{2+o(1)}$. (We in fact prove a more intricate complexity bound that is slightly better.) The best previous general algorithm had (deterministic) complexity exponential in $k$. We also present some experimental data evincing the potential practicality of our algorithm.

math.NT↗
arXiv2019

A Faster Solution to Smale's 17th Problem I: Real Binomial Systems

Suppose $F:=(f_1,\ldots,f_n)$ is a system of random $n$-variate polynomials with $f_i$ having degree $\leq\!d_i$ and the coefficient of $x^{a_1}_1\cdots x^{a_n}_n$ in $f_i$ being an independent complex Gaussian of mean $0$ and variance $\frac{d_i!}{a_1!\cdots a_n!\left(d_i-\sum^n_{j=1}a_j \right)!}$. Recent progress on Smale's 17th Problem by Lairez --- building upon seminal work of Shub, Beltran, Pardo, Bürgisser, and Cucker --- has resulted in a deterministic algorithm that finds a single (complex) approximate root of $F$ using just $N^{O(1)}$ arithmetic operations on average, where $N\!:=\!\sum^n_{i=1}\frac{(n+d_i)!}{n!d_i!}$ ($=n(n+\max_i d_i)^{O(\min\{n,\max_i d_i)\}}$) is the maximum possible total number of monomial terms for such an $F$. However, can one go faster when the number of terms is smaller, and we restrict to real coefficient and real roots? And can one still maintain average-case polynomial-time with more general probability measures? We show the answer is yes when $F$ is instead a binomial system --- a case whose numerical solution is a key step in polyhedral homotopy algorithms for solving arbitrary polynomial systems. We give a deterministic algorithm that finds a real approximate root (or correctly decides there are none) using just $O(n^2(\log(n)+\log\max_i d_i))$ arithmetic operations on average. Furthermore, our approach allows Gaussians with arbitrary variance. We also discuss briefly the obstructions to maintaining average-case time polynomial in $n\log \max_i d_i$ when $F$ has more terms.

math.AG↗
arXiv2018

Dedekind Zeta Zeroes and Faster Complex Dimension Computation

Thanks to earlier work of Koiran, it is known that the truth of the Generalized Riemann Hypothesis (GRH) implies that the dimension of algebraic sets over the complex numbers can be determined within the polynomial-hierarchy. The truth of GRH thus provides a direct connection between a concrete algebraic geometry problem and the P vs.NP Problem, in a radically different direction from the geometric complexity theory approach to VP vs. VNP. We explore more plausible hypotheses yielding the same speed-up. One minimalist hypothesis we derive involves improving the error term (as a function of the degree, coefficient height, and $x$) on the fraction of primes $p\!\leq\!x$ for which a univariate polynomial has roots mod $p$. A second minimalist hypothesis involves sharpening current zero-free regions for Dedekind zeta functions. Both our hypotheses allow failures of GRH but still enable complex dimension computation in the polynomial hierarchy.

cs.CC↗
arXiv2017

Counting Roots of Polynomials Over Prime Power Rings

Suppose $p$ is a prime, $t$ is a positive integer, and $f\!\in\!\mathbb{Z}[x]$ is a univariate polynomial of degree $d$ with coefficients of absolute value $<\!p^t$. We show that for any fixed $t$, we can compute the number of roots in $\mathbb{Z}/(p^t)$ of $f$ in deterministic time $(d+\log p)^{O(1)}$. This fixed parameter tractability appears to be new for $t\!\geq\!3$. A consequence for arithmetic geometry is that we can efficiently compute Igusa zeta functions $Z$, for univariate polynomials, assuming the degree of $Z$ is fixed.

math.NT↗
arXiv2017

A-Discriminants for Complex Exponents, and Counting Real Isotopy Types

We extend the definition of $\mathcal{A}$-discriminant varieties, and Kapranov's parametrization of $\mathcal{A}$-discriminant varieties, to complex exponents. As an application, we study the special case where $\mathcal{A}$ is a fixed real $n\times (n+3)$ matrix whose columns form the spectrum of an $n$-variate exponential sum $g$ with fixed sign vector for its coefficients: We prove that the number of possible isotopy types for the real zero set of $g$ is $O(n^2)$. The best previous upper bound was $2^{O(n^4)}$. Along the way, we also show that the singular loci of our generalized $\mathcal{A}$-discriminants are images of low-degree algebraic sets under certain analytic maps.

math.AG↗
arXiv2017

New Subexponential Fewnomial Hypersurface Bounds

Suppose $c_1,\ldots,c_{n+k}$ are real numbers, $\{a_1,\ldots,a_{n+k}\}\!\subset\!\mathbb{R}^n$ is a set of points not all lying in the same affine hyperplane, $y\!\in\!\mathbb{R}^n$, $a_j\cdot y$ denotes the standard real inner product of $a_j$ and $y$, and we set $g(y)\!:=\!\sum^{n+k}_{j=1} c_j e^{a_j\cdot y}$. We prove that, for generic $c_j$, the number of connected components of the real zero set of $g$ is $O\!\left(n^2+\sqrt{2}^{k^2}(n+2)^{k-2}\right)$. The best previous upper bounds, when restricted to the special case $k\!=\!3$ and counting just the non-compact components, were already exponential in $n$.

math.AG↗