-
New Lower Bounds against Homogeneous Non-Commutative Circuits
Authors:
Prerona Chatterjee,
Pavel Hrubeš
Abstract:
We give several new lower bounds on size of homogeneous non-commutative circuits. We present an explicit homogeneous bivariate polynomial of degree $d$ which requires homogeneous non-commutative circuit of size $Ω(d/\log d)$. For an $n$-variate polynomial with $n>1$, the result can be improved to $Ω(nd)$, if $d\leq n$, or $Ω(nd \frac{\log n}{\log d})$, if $d\geq n$.
Under the same assumptions, w…
▽ More
We give several new lower bounds on size of homogeneous non-commutative circuits. We present an explicit homogeneous bivariate polynomial of degree $d$ which requires homogeneous non-commutative circuit of size $Ω(d/\log d)$. For an $n$-variate polynomial with $n>1$, the result can be improved to $Ω(nd)$, if $d\leq n$, or $Ω(nd \frac{\log n}{\log d})$, if $d\geq n$.
Under the same assumptions, we also give a quadratic lower bound for the ordered version of the central symmetric polynomial.
△ Less
Submitted 4 January, 2023;
originally announced January 2023.
-
On the extension complexity of polytopes separating subsets of the Boolean cube
Authors:
Pavel Hrubeš,
Navid Talebanfard
Abstract:
We show that
1. for every $A\subseteq \{0, 1\}^n$, there exists a polytope $P\subseteq \mathbb{R}^n$ with $P \cap \{0, 1\}^n = A$ and extension complexity $O(2^{n/2})$,
2. there exists an $A\subseteq \{0, 1\}^n$ such that the extension complexity of any $P$ with $P\cap \{0, 1\}^n = A$ must be at least $2^{\frac{n}{3}(1-o(1))}$.
We also remark that the extension complexity of any 0/1-polytope…
▽ More
We show that
1. for every $A\subseteq \{0, 1\}^n$, there exists a polytope $P\subseteq \mathbb{R}^n$ with $P \cap \{0, 1\}^n = A$ and extension complexity $O(2^{n/2})$,
2. there exists an $A\subseteq \{0, 1\}^n$ such that the extension complexity of any $P$ with $P\cap \{0, 1\}^n = A$ must be at least $2^{\frac{n}{3}(1-o(1))}$.
We also remark that the extension complexity of any 0/1-polytope in $\mathbb{R}^n$ is at most $O(2^n/n)$ and pose the problem whether the upper bound can be improved to $O(2^{cn})$, for $c<1$.
△ Less
Submitted 25 May, 2021;
originally announced May 2021.
-
On the distribution of runners on a circle
Authors:
Pavel Hrubes
Abstract:
Consider $n$ runners running on a circular track of unit length with constant speeds such that $k$ of the speeds are distinct. We show that, at some time, there will exist a sector $S$ which contains at least $|S|n+ Ω(\sqrt{k})$ runners. The result can be generalized as follows. Let $f(x,y)$ be a complex bivariate polynomial whose Newton polytope has $k$ vertices. Then there exists…
▽ More
Consider $n$ runners running on a circular track of unit length with constant speeds such that $k$ of the speeds are distinct. We show that, at some time, there will exist a sector $S$ which contains at least $|S|n+ Ω(\sqrt{k})$ runners. The result can be generalized as follows. Let $f(x,y)$ be a complex bivariate polynomial whose Newton polytope has $k$ vertices. Then there exists $a\in {\mathbb C}\setminus\{0\}$ and a complex sector $S=\{re^{\imath θ}: r>0, α\leq θ\leq β\}$ such that the univariate polynomial $f(x,a)$ contains at least $\frac{β-α}{2π}n+Ω(\sqrt{k})$ non-zero roots in $S$ (where $n$ is the total number of such roots and $0\leq (β-α)\leq 2π$). This shows that the Real $τ$-Conjecture of Koiran implies the conjecture on Newton polytopes of Koiran et al.
△ Less
Submitted 5 April, 2020; v1 submitted 6 June, 2019;
originally announced June 2019.
-
A learning problem that is independent of the set theory ZFC axioms
Authors:
Shai Ben-David,
Pavel Hrubes,
Shay Moran,
Amir Shpilka,
Amir Yehudayoff
Abstract:
We consider the following statistical estimation problem: given a family F of real valued functions over some domain X and an i.i.d. sample drawn from an unknown distribution P over X, find h in F such that the expectation of h w.r.t. P is probably approximately equal to the supremum over expectations on members of F. This Expectation Maximization (EMX) problem captures many well studied learning…
▽ More
We consider the following statistical estimation problem: given a family F of real valued functions over some domain X and an i.i.d. sample drawn from an unknown distribution P over X, find h in F such that the expectation of h w.r.t. P is probably approximately equal to the supremum over expectations on members of F. This Expectation Maximization (EMX) problem captures many well studied learning problems; in fact, it is equivalent to Vapnik's general setting of learning.
Surprisingly, we show that the EMX learnability, as well as the learning rates of some basic class F, depend on the cardinality of the continuum and is therefore independent of the set theory ZFC axioms (that are widely accepted as a formalization of the notion of a mathematical proof).
We focus on the case where the functions in F are Boolean, which generalizes classification problems. We study the interaction between the statistical sample complexity of F and its combinatorial structure. We introduce a new version of sample compression schemes and show that it characterizes EMX learnability for a wide family of classes. However, we show that for the class of finite subsets of the real line, the existence of such compression schemes is independent of set theory. We conclude that the learnability of that class with respect to the family of probability distributions of countable support is independent of the set theory ZFC axioms.
We also explore the existence of a "VC-dimension-like" parameter that captures learnability in this setting. Our results imply that that there exist no "finitary" combinatorial parameter that characterizes EMX learnability in a way similar to the VC-dimension based characterization of binary valued classification problems.
△ Less
Submitted 14 November, 2017;
originally announced November 2017.
-
On families of anticommuting matrices
Authors:
Pavel Hrubeš
Abstract:
Let $e_{1},\dots, e_{k}$ be complex $n\times n$ matrices such that $e_{i}e_{j}=-e_{j}e_{i}$ whenever $i\not=j$. We conjecture that $\hbox{rk}(e_{1}^{2})+\hbox{rk}(e_{2}^{2})+\cdots+\hbox{rk}(e_{k}^{2})\leq O(n\log n)$, and prove some results in this direction.
Let $e_{1},\dots, e_{k}$ be complex $n\times n$ matrices such that $e_{i}e_{j}=-e_{j}e_{i}$ whenever $i\not=j$. We conjecture that $\hbox{rk}(e_{1}^{2})+\hbox{rk}(e_{2}^{2})+\cdots+\hbox{rk}(e_{k}^{2})\leq O(n\log n)$, and prove some results in this direction.
△ Less
Submitted 18 December, 2014;
originally announced December 2014.
-
Short Proofs for the Determinant Identities
Authors:
Pavel Hrubes,
Iddo Tzameret
Abstract:
We study arithmetic proof systems P_c(F) and P_f(F) operating with arithmetic circuits and arithmetic formulas, respectively, that prove polynomial identities over a field F. We establish a series of structural theorems about these proof systems, the main one stating that P_c(F) proofs can be balanced: if a polynomial identity of syntactic degree d and depth k has a P_c(F) proof of size s, then it…
▽ More
We study arithmetic proof systems P_c(F) and P_f(F) operating with arithmetic circuits and arithmetic formulas, respectively, that prove polynomial identities over a field F. We establish a series of structural theorems about these proof systems, the main one stating that P_c(F) proofs can be balanced: if a polynomial identity of syntactic degree d and depth k has a P_c(F) proof of size s, then it also has a P_c(F) proof of size poly(s,d) and depth O(k+\log^2 d + \log d\cd \log s). As a corollary, we obtain a quasipolynomial simulation of P_c(F) by P_f(F), for identities of a polynomial syntactic degree.
Using these results we obtain the following: consider the identities det(XY) = det(X)det(Y) and det(Z)= z_{11}... z_{nn}, where X,Y and Z are nxn square matrices and Z is a triangular matrix with z_{11},..., z_{nn} on the diagonal (and det is the determinant polynomial). Then we can construct a polynomial-size arithmetic circuit det such that the above identities have P_c(F) proofs of polynomial-size and O(\log^2 n) depth. Moreover, there exists an arithmetic formula det of size n^{O(\log n)} such that the above identities have P_f(F) proofs of size n^{O(\log n)}.
This yields a solution to a basic open problem in propositional proof complexity, namely, whether there are polynomial-size NC^2-Frege proofs for the determinant identities and the hard matrix identities, as considered, e.g., in Soltys and Cook (2004) (cf., Beame and Pitassi (1998)). We show that matrix identities like AB=I {\to} BA=I (for matrices over the two element field) as well as basic properties of the determinant have polynomial-size NC^2-Frege proofs, and quasipolynomial-size Frege proofs.
△ Less
Submitted 22 April, 2013; v1 submitted 29 December, 2011;
originally announced December 2011.
-
Homogeneous formulas and symmetric polynomials
Authors:
Pavel Hrubes,
Amir Yehudayoff
Abstract:
We investigate the arithmetic formula complexity of the elementary symmetric polynomials S(k,n). We show that every multilinear homogeneous formula computing S(k,n) has size at least k^(Omega(log k))n, and that product-depth d multilinear homogeneous formulas for S(k,n) have size at least 2^(Omega(k^{1/d}))n. Since S(n,2n) has a multilinear formula of size O(n^2), we obtain a superpolynomial sep…
▽ More
We investigate the arithmetic formula complexity of the elementary symmetric polynomials S(k,n). We show that every multilinear homogeneous formula computing S(k,n) has size at least k^(Omega(log k))n, and that product-depth d multilinear homogeneous formulas for S(k,n) have size at least 2^(Omega(k^{1/d}))n. Since S(n,2n) has a multilinear formula of size O(n^2), we obtain a superpolynomial separation between multilinear and multilinear homogeneous formulas. We also show that S(k,n) can be computed by homogeneous formulas of size k^(O(log k))n, answering a question of Nisan and Wigderson. Finally, we present a superpolynomial separation between monotone and non-monotone formulas in the noncommutative setting, answering a question of Nisan.
△ Less
Submitted 15 July, 2009;
originally announced July 2009.
-
On hierarchies of universal predicates
Authors:
Pavel Hrubes
Abstract:
We investigate a hierarchy of arithmetical structures obtained by a transfinite addition of a canonic universal predicate, where the canonic universal predicate for M is defined as a minimum universal predicate for M in terms of definability. We determine the upper bound of the hierarchy and give a characterisation for the sets definable in the hierarchy.
We investigate a hierarchy of arithmetical structures obtained by a transfinite addition of a canonic universal predicate, where the canonic universal predicate for M is defined as a minimum universal predicate for M in terms of definability. We determine the upper bound of the hierarchy and give a characterisation for the sets definable in the hierarchy.
△ Less
Submitted 24 March, 2007;
originally announced March 2007.
