arXiv subjects

Mrinal Kumar

Publications and source records attributed to Mrinal Kumar.

At least 55 records · Page 3Linked to original sources

arXiv2020

A Warm Start Method for Solving Chance Constrained Optimal Control Problems

A warm start method is developed for efficiently solving complex chance constrained optimal control problems. The warm start method addresses the computational challenges of solving chance constrained optimal control problems using biased kernel density estimators and Legendre-Gauss-Radau collocation with an $hp$ adaptive mesh refinement method. To address the computational challenges, the warm start method improves both the starting point for the chance constrained optimal control problem, as well as the efficiency of cycling through mesh refinement iterations. The improvement is accomplished by tuning a parameter of the kernel density estimator, as well as implementing a kernel switch as part of the solution process. Additionally, the number of samples for the biased kernel density estimator is set to incrementally increase through a series of mesh refinement iterations. Thus, the warm start method is a combination of tuning a parameter, a kernel switch, and an incremental increase in sample size. This warm start method is successfully applied to solve two challenging chance constrained optimal control problems in a computationally efficient manner using biased kernel density estimators and Legendre-Gauss-Radau collocation.

math.OC↗

arXiv2020

Derandomization from Algebraic Hardness

A hitting-set generator (HSG) is a polynomial map $G:\mathbb{F}^k \to \mathbb{F}^n$ such that for all $n$-variate polynomials $C$ of small enough circuit size and degree, if $C$ is nonzero, then $C\circ G$ is nonzero. In this paper, we give a new construction of such an HSG assuming that we have an explicit polynomial of sufficient hardness. Formally, we prove the following over any field of characteristic zero: Let $k\in \mathbb{N}$ and $δ> 0$ be arbitrary constants. Suppose $\{P_d\}_{d\in \mathbb{N}}$ is an explicit family of $k$-variate polynomials such that $\operatorname{deg} P_d = d$ and $P_d$ requires algebraic circuits of size $d^δ$. Then, there are explicit hitting sets of polynomial size for $\mathsf{VP}$. This is the first HSG in the algebraic setting that yields a complete derandomization of polynomial identity testing (PIT) for general circuits from a suitable algebraic hardness assumption. As a direct consequence, we show that even saving a single point from the "trivial" explicit, exponential sized hitting sets for constant-variate polynomials of low individual degree which are computable by small circuits, implies a deterministic polynomial time algorithm for PIT. More precisely, we show the following: Let $k\in \mathbb{N}$ and $δ> 0$ be arbitrary constants. Suppose for every $s$ large enough, there is an explicit hitting set of size at most $((s+1)^k - 1)$ for the class of $k$-variate polynomials of individual degree $s$ that are computable by size $s^δ$ circuits. Then there is an explicit hitting set of size $\operatorname{poly}(s)$ for the class of $s$-variate polynomials, of degree $s$, that are computable by size $s$ circuits. As a consequence, we give a deterministic polynomial time construction of hitting sets for algebraic circuits, if a strengthening of the $τ$-Conjecture of Shub and Smale is true.

cs.CC↗

arXiv2020

Method for Chance Constrained Optimal Control Using Biased Kernel Density Estimators

A method is developed to numerically solve chance constrained optimal control problems. The chance constraints are reformulated as nonlinear constraints that retain the probability properties of the original constraint. The reformulation transforms the chance constrained optimal control problem into a deterministic optimal control problem that can be solved numerically. The new method developed in this paper approximates the chance constraints using Markov Chain Monte Carlo (MCMC) sampling and kernel density estimators whose kernels have integral functions that bound the indicator function. The nonlinear constraints resulting from the application of kernel density estimators are designed with bounds that do not violate the bounds of the original chance constraint. The method is tested on a non-trivial chance constrained modification of a soft lunar landing optimal control problem and the results are compared with results obtained using a conservative deterministic formulation of the optimal control problem. The results show that this new method efficiently solves chance constrained optimal control problems.

math.OC↗

arXiv2020

A Quadratic Lower Bound for Algebraic Branching Programs and Formulas

We show that any Algebraic Branching Program (ABP) computing the polynomial $\sum_{i = 1}^n x_i^n$ has at least $Ω(n^2)$ vertices. This improves upon the lower bound of $Ω(n\log n)$, which follows from the classical result of Baur and Strassen [Str73, BS83], and extends the results in [K19], which showed a quadratic lower bound for \emph{homogeneous} ABPs computing the same polynomial. Our proof relies on a notion of depth reduction which is reminiscent of similar statements in the context of matrix rigidity, and shows that any small enough ABP computing the polynomial $\sum_{i=1}^n x_i^n$ can be depth reduced to essentially a homogeneous ABP of the same size which computes the polynomial $\sum_{i = 1}^n x_i^n + ε(x_1, \ldots, x_n)$, for a structured "error polynomial" $ε(x_1, \ldots, x_n)$. To complete the proof, we then observe that the lower bound in [K19] is robust enough and continues to hold for all polynomials $\sum_{i = 1}^n x_i^n + ε(x_1, \ldots, x_n)$, where $ε(x_1, \ldots, x_n)$ has the appropriate structure. We also use our ideas to show an $Ω(n^2)$ lower bound of the size of algebraic formulas computing the elementary symmetric polynomial of degree $0.1n$ on $n$ variables. This is a slight improvement upon the prior best known formula lower bound (proved for a different polynomial) of $Ω(n^2/\log n)$ [Nec66, K85, SY10]. Interestingly, this lower bound is asymptotically better than $n^2/\log n$, the strongest lower bound that can be proved using previous methods. This lower bound also matches the upper bound, due to Ben-Or, who showed that elementary symmetric polynomials can be computed by algebraic formula (in fact depth-$3$ formula) of size $O(n^2)$. Prior to this work, Ben-Or's construction was known to be optimal only for algebraic formulas of depth-$3$ [SW01].

cs.CC↗

arXiv2019

Schur Polynomials do not have small formulas if the Determinant doesn't!

Schur Polynomials are families of symmetric polynomials that have been classically studied in Combinatorics and Algebra alike. They play a central role in the study of Symmetric functions, in Representation theory [Sta99], in Schubert calculus [LM10] as well as in Enumerative combinatorics [Gas96, Sta84, Sta99]. In recent years, they have also shown up in various incarnations in Computer Science, e.g, Quantum computation [HRTS00, OW15] and Geometric complexity theory [IP17]. However, unlike some other families of symmetric polynomials like the Elementary Symmetric polynomials, the Power Symmetric polynomials and the Complete Homogeneous Symmetric polynomials, the computational complexity of syntactically computing Schur polynomials has not been studied much. In particular, it is not known whether Schur polynomials can be computed efficiently by algebraic formulas. In this work, we address this question, and show that unless \emph{every} polynomial with a small algebraic branching program (ABP) has a small algebraic formula, there are Schur polynomials that cannot be computed by algebraic formula of polynomial size. In other words, unless the algebraic complexity class $\mathrm{VBP}$ is equal to the complexity class $\mathrm{VF}$, there exist Schur polynomials which do not have polynomial size algebraic formulas. As a consequence of our proof, we also show that computing the determinant of certain \emph{generalized} Vandermonde matrices is essentially as hard as computing the general symbolic determinant. To the best of our knowledge, these are one of the first hardness results of this kind for families of polynomials which are not \emph{multilinear}. A key ingredient of our proof is the study of composition of \emph{well behaved} algebraically independent polynomials with a homogeneous polynomial, and might be of independent interest.

cs.CC↗

arXiv2019

Closure of VP under taking factors: a short and simple proof

In this note, we give a short, simple and almost completely self contained proof of a classical result of Kaltofen [Kal86, Kal87, Kal89] which shows that if an $n$ variate degree $d$ polynomial $f$ can be computed by an arithmetic circuit of size $s$, then each of its factors can be computed by an arithmetic circuit of size at most $\textsf{poly}\left(s, n, d\right)$. However, unlike Kaltofen's argument, our proof does not directly give an efficient algorithm for computing the circuits for the factors of $f$.

cs.CC↗

arXiv2019

Lower Bounds for Matrix Factorization

We study the problem of constructing explicit families of matrices which cannot be expressed as a product of a few sparse matrices. In addition to being a natural mathematical question on its own, this problem appears in various incarnations in computer science; the most significant being in the context of lower bounds for algebraic circuits which compute linear transformations, matrix rigidity and data structure lower bounds. We first show, for every constant $d$, a deterministic construction in subexponential time of a family $\{M_n\}$ of $n \times n$ matrices which cannot be expressed as a product $M_n = A_1 \cdots A_d$ where the total sparsity of $A_1,\ldots,A_d$ is less than $n^{1+1/(2d)}$. In other words, any depth-$d$ linear circuit computing the linear transformation $M_n\cdot x$ has size at least $n^{1+Ω(1/d)}$. This improves upon the prior best lower bounds for this problem, which are barely super-linear, and were obtained by a long line of research based on the study of super-concentrators (albeit at the cost of a blow up in the time required to construct these matrices). We then outline an approach for proving improved lower bounds through a certain derandomization problem, and use this approach to prove asymptotically optimal quadratic lower bounds for natural special cases, which generalize many of the common matrix decompositions.

cs.CC↗

arXiv2019

Towards Optimal Depth Reductions for Syntactically Multilinear Circuits

We show that any $n$-variate polynomial computable by a syntactically multilinear circuit of size $\operatorname{poly}(n)$ can be computed by a depth-$4$ syntactically multilinear ($ΣΠΣΠ$) circuit of size at most $\exp\left({O\left(\sqrt{n\log n}\right)}\right)$. For degree $d = ω(n/\log n)$, this improves upon the upper bound of $\exp\left({O(\sqrt{d}\log n)}\right)$ obtained by Tavenas~\cite{T15} for general circuits, and is known to be asymptotically optimal in the exponent when $d < n^ε$ for a small enough constant $ε$. Our upper bound matches the lower bound of $\exp\left({Ω\left(\sqrt{n\log n}\right)}\right)$ proved by Raz and Yehudayoff~\cite{RY09}, and thus cannot be improved further in the exponent. Our results hold over all fields and also generalize to circuits of small individual degree. More generally, we show that an $n$-variate polynomial computable by a syntactically multilinear circuit of size $\operatorname{poly}(n)$ can be computed by a syntactically multilinear circuit of product-depth $Δ$ of size at most $\exp\left(O\left(Δ\cdot (n/\log n)^{1/Δ} \cdot \log n\right)\right)$. It follows from the lower bounds of Raz and Yehudayoff (CC 2009) that in general, for constant $Δ$, the exponent in this upper bound is tight and cannot be improved to $o\left(\left(n/\log n\right)^{1/Δ}\cdot \log n\right)$.

cs.CC↗

arXiv2018

Arithmetic Circuits with Locally Low Algebraic Rank

In recent years, there has been a flurry of activity towards proving lower bounds for homogeneous depth-4 arithmetic circuits, which has brought us very close to statements that are known to imply $\textsf{VP} \neq \textsf{VNP}$. It is open if these techniques can go beyond homogeneity, and in this paper we make some progress in this direction by considering depth-4 circuits of low algebraic rank, which are a natural extension of homogeneous depth-4 circuits. A depth-4 circuit is a representation of an $N$-variate, degree-$n$ polynomial $P$ as \[ P = \sum_{i = 1}^T Q_{i1}\cdot Q_{i2}\cdot \cdots \cdot Q_{it} \; , \] where the $Q_{ij}$ are given by their monomial expansion. Homogeneity adds the constraint that for every $i \in [T]$, $\sum_{j} \operatorname{deg}(Q_{ij}) = n$. We study an extension, where, for every $i \in [T]$, the algebraic rank of the set $\{Q_{i1}, Q_{i2}, \ldots ,Q_{it}\}$ of polynomials is at most some parameter $k$. Already for $k = n$, these circuits are a generalization of the class of homogeneous depth-4 circuits, where in particular $t \leq n$ (and hence $k \leq n$). We study lower bounds and polynomial identity tests for such circuits and prove the following results. We show an $\exp{(Ω(\sqrt{n}\log N))}$ lower bound for such circuits for an explicit $N$ variate degree $n$ polynomial family when $k \leq n$. We also show quasipolynomial hitting sets when the degree of each $Q_{ij}$ and the $k$ are at most $\operatorname{poly}(\log n)$. A key technical ingredient of the proofs, which may be of independent interest, is a result which states that over any field of characteristic zero, up to a translation, every polynomial in a set of polynomials can be written as a function of the polynomials in a transcendence basis of the set. We combine this with methods based on shifted partial derivatives to obtain our final results.

cs.CC↗

arXiv2018

On Multilinear Forms: Bias, Correlation, and Tensor Rank

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 $f(X_1,X_2, \dots, X_d) : \left(GF(2)^{k}\right)^d \rightarrow GF(2)$ has correlation $2^{-k(1-o(1))}$ with any polynomial of degree at most $d/10$. This result is proved by giving near-optimal bounds on the bias of random $d$-linear form, which is in turn proved by giving near-optimal bounds on the probability that a random rank-$t$ $d$-linear form is identically zero. 2. Tensor-rank vs Bias : We show that if a $d$-dimensional tensor has small rank, then the bias of the associated $d$-linear form is large. More precisely, given any $d$-dimensional tensor $$T :\underbrace{[k]\times \ldots [k]}_{\text{$d$ times}}\to GF(2)$$ of rank at most $t$, the bias of the associated $d$-linear form $$f_T(X_1,\ldots,X_d) := \sum_{(i_1,\dots,i_d) \in [k]^d} T(i_1,i_2,\ldots, i_d) X_{1,i_1}\cdot X_{1,i_2}\cdots X_{d,i_d}$$ is at least $\left(1-\frac1{2^{d-1}}\right)^t$. The above bias vs tensor-rank connection suggests a natural approach to proving nontrivial tensor-rank lower bounds for $d=3$. In particular, we use this approach to prove that the finite field multiplication tensor has tensor rank at least $3.52 k$ matching the best known lower bound for any explicit tensor in three dimensions over $GF(2)$.

cs.CC↗

arXiv2018

On top fan-in vs formal degree for depth-$3$ arithmetic circuits

We show that over the field of complex numbers, \emph{every} homogeneous polynomial of degree $d$ can be approximated (in the border complexity sense) by a depth-$3$ arithmetic circuit of top fan-in at most $d+1$. This is quite surprising since there exist homogeneous polynomials $P$ on $n$ variables of degree $2$, such that any depth-$3$ arithmetic circuit computing $P$ must have top fan-in at least $Ω(n)$. As an application, we get a new tradeoff between the top fan-in and formal degree in an approximate analog of the celebrated depth reduction result of Gupta, Kamath, Kayal and Saptharishi [GKKS13]. Formally, we show that if a degree $d$ homogeneous polynomial $P$ can be computed by an arithmetic circuit of size $s$, then for every $t \leq d$, $P$ is in the border of a depth-$3$ circuit of top fan-in $s^{O(t)}$ and formal degree $s^{O(d/t)}$. To the best of our knowledge, the upper bound on the top fan-in in the original proof of [GKKS13] is always at least $s^{O(\sqrt{d})}$, regardless of the formal degree.

cs.CC↗

arXiv2018

Some Closure Results for Polynomial Factorization and Applications

In a sequence of seminal results in the 80's, Kaltofen showed that the complexity class VP is closed under taking factors. A natural question in this context is to understand if other natural classes of multivariate polynomials, for instance, arithmetic formulas, algebraic branching programs, bounded depth arithmetic circuits or the class VNP, are closed under taking factors. In this paper, we show that all factors of degree at most $\log^a n$ of polynomials with poly(n) size depth $k$ circuits have poly(n) size circuits of depth at most $O(k + a)$. This partially answers a question of Shpilka-Yehudayoff and has applications to hardness-randomness tradeoffs for bounded depth arithmetic circuits. More precisely, this shows that a superpolynomial lower bound for bounded depth arithmetic circuits, for a family of explicit polynomials of degree poly$(\log n)$ implies deterministic sub-exponential time algorithms for polynomial identity testing (PIT) for bounded depth arithmetic circuits. This is incomparable to a beautiful result of Dvir et al., where they showed that super-polynomial lower bounds for constant depth arithmetic circuits for any explicit family of polynomials (of potentially high degree) implies sub-exponential time deterministic PIT for bounded depth circuits of bounded individual degree. Thus, we remove the "bounded individual degree" condition in [DSY09] at the cost of strengthening the hardness assumption to hold for polynomials of low degree. As direct applications of our techniques, we also show that the complexity class VNP is closed under taking factors, thereby confirming a conjecture of Bürgisser and get an alternate proof of the fact (first shown by Dutta et al.) that if a polynomial $Q$ of degree at most $d$ divides a polynomial $P$ computable by a formula of size $s$, then $Q$ has a formula of size at most poly$(s, d^{\log d}, deg(P))$.

cs.CC↗

arXiv2017

Unbalancing Sets and an Almost Quadratic Lower Bound for Syntactically Multilinear Arithmetic Circuits

We prove a lower bound of $Ω(n^2/\log^2 n)$ on the size of any syntactically multilinear arithmetic circuit computing some explicit multilinear polynomial $f(x_1, \ldots, x_n)$. Our approach expands and improves upon a result of Raz, Shpilka and Yehudayoff ([RSY08]), who proved a lower bound of $Ω(n^{4/3}/\log^2 n)$ for the same polynomial. Our improvement follows from an asymptotically optimal lower bound for a generalized version of Galvin's problem in extremal set theory.

cs.CC↗

arXiv2017

The Chasm at Depth Four, and Tensor Rank : Old results, new insights

Agrawal and Vinay [AV08] showed how any polynomial size arithmetic circuit can be thought of as a depth four arithmetic circuit of subexponential size. The resulting circuit size in this simulation was more carefully analyzed by Korian [Koiran] and subsequently by Tavenas [Tav13]. We provide a simple proof of this chain of results. We then abstract the main ingredient to apply it to formulas and constant depth circuits, and show more structured depth reductions for them. In an apriori surprising result, Raz [Raz10] showed that for any $n$ and $d$, such that $ ω(1) \leq d \leq O\left(\frac{\log n}{\log\log n}\right)$, constructing explicit tensors $T:[n]^d \rightarrow F$ of high enough rank would imply superpolynomial lower bounds for arithmetic formulas over the field $F$. Using the additional structure we obtain from our proof of the depth reduction for arithmetic formulas, we give a new and arguably simpler proof of this connection. We also extend this result for homogeneous formulas to show that, in fact, the connection holds for any $d$ such that $ω(1) \leq d \leq n^{o(1)}$.

cs.CC↗

arXiv2017

Towards an algebraic natural proofs barrier via polynomial identity testing

We observe that a certain kind of algebraic proof - which covers essentially all known algebraic circuit lower bounds to date - cannot be used to prove lower bounds against VP if and only if what we call succinct hitting sets exist for VP. This is analogous to the Razborov-Rudich natural proofs barrier in Boolean circuit complexity, in that we rule out a large class of lower bound techniques under a derandomization assumption. We also discuss connections between this algebraic natural proofs barrier, geometric complexity theory, and (algebraic) proof complexity.

cs.CC↗

arXiv2016

Functional lower bounds for arithmetic circuits and connections to boolean circuit complexity

We say that a circuit $C$ over a field $F$ functionally computes an $n$-variate polynomial $P$ if for every $x \in \{0,1\}^n$ we have that $C(x) = P(x)$. This is in contrast to syntactically computing $P$, when $C \equiv P$ as formal polynomials. In this paper, we study the question of proving lower bounds for homogeneous depth-$3$ and depth-$4$ arithmetic circuits for functional computation. We prove the following results : 1. Exponential lower bounds homogeneous depth-$3$ arithmetic circuits for a polynomial in $VNP$. 2. Exponential lower bounds for homogeneous depth-$4$ arithmetic circuits with bounded individual degree for a polynomial in $VNP$. Our main motivation for this line of research comes from our observation that strong enough functional lower bounds for even very special depth-$4$ arithmetic circuits for the Permanent imply a separation between ${\#}P$ and $ACC$. Thus, improving the second result to get rid of the bounded individual degree condition could lead to substantial progress in boolean circuit complexity. Besides, it is known from a recent result of Kumar and Saptharishi [KS15] that over constant sized finite fields, strong enough average case functional lower bounds for homogeneous depth-$4$ circuits imply superpolynomial lower bounds for homogeneous depth-$5$ circuits. Our proofs are based on a family of new complexity measures called shifted evaluation dimension, and might be of independent interest.

cs.CC↗

arXiv2015

An exponential lower bound for homogeneous depth-5 circuits over finite fields

In this paper, we show exponential lower bounds for the class of homogeneous depth-$5$ circuits over all small finite fields. More formally, we show that there is an explicit family $\{P_d : d \in \mathbb{N}\}$ of polynomials in $\mathsf{VNP}$, where $P_d$ is of degree $d$ in $n = d^{O(1)}$ variables, such that over all finite fields $\mathbb{F}_q$, any homogeneous depth-$5$ circuit which computes $P_d$ must have size at least $\exp(Ω_q(\sqrt{d}))$. To the best of our knowledge, this is the first super-polynomial lower bound for this class for any field $\mathbb{F}_q \neq \mathbb{F}_2$. Our proof builds up on the ideas developed on the way to proving lower bounds for homogeneous depth-$4$ circuits [GKKS13, FLMS13, KLSS14, KS14] and for non-homogeneous depth-$3$ circuits over finite fields [GK98, GR00]. Our key insight is to look at the space of shifted partial derivatives of a polynomial as a space of functions from $\mathbb{F}_q^n \rightarrow \mathbb{F}_q$ as opposed to looking at them as a space of formal polynomials and builds over a tighter analysis of the lower bound of Kumar and Saraf [KS14].

cs.CC↗

arXiv2015

Sums of products of polynomials in few variables : lower bounds and polynomial identity testing

We study the complexity of representing polynomials as a sum of products of polynomials in few variables. More precisely, we study representations of the form $$P = \sum_{i = 1}^T \prod_{j = 1}^d Q_{ij}$$ such that each $Q_{ij}$ is an arbitrary polynomial that depends on at most $s$ variables. We prove the following results. 1. Over fields of characteristic zero, for every constant $μ$ such that $0 \leq μ< 1$, we give an explicit family of polynomials $\{P_{N}\}$, where $P_{N}$ is of degree $n$ in $N = n^{O(1)}$ variables, such that any representation of the above type for $P_{N}$ with $s = N^μ$ requires $Td \geq n^{Ω(\sqrt{n})}$. This strengthens a recent result of Kayal and Saha [KS14a] which showed similar lower bounds for the model of sums of products of linear forms in few variables. It is known that any asymptotic improvement in the exponent of the lower bounds (even for $s = \sqrt{n}$) would separate VP and VNP[KS14a]. 2. We obtain a deterministic subexponential time blackbox polynomial identity testing (PIT) algorithm for circuits computed by the above model when $T$ and the individual degree of each variable in $P$ are at most $\log^{O(1)} N$ and $s \leq N^μ$ for any constant $μ< 1/2$. We get quasipolynomial running time when $s < \log^{O(1)} N$. The PIT algorithm is obtained by combining our lower bounds with the hardness-randomness tradeoffs developed in [DSY09, KI04]. To the best of our knowledge, this is the first nontrivial PIT algorithm for this model (even for the case $s=2$), and the first nontrivial PIT algorithm obtained from lower bounds for small depth circuits.

cs.CC↗