arXiv subjects

Jesse Goodman

Publications and source records attributed to Jesse Goodman.

At least 19 recordsLinked to original sources

arXiv2026

A general framework for computation and estimation using the saddlepoint approximation

The saddlepoint approximation provides highly accurate approximations to probability density and mass functions using only the corresponding moment generating functions (MGFs). Recent work has increasingly seen the saddlepoint approximation applied to likelihood functions, enabling likelihood-based inference in models where exact likelihoods are intractable. However, existing implementations have largely been developed on a model-by-model basis, and the methodology remains underutilized because of the conceptual and computational challenges of working with MGFs. We introduce a unified framework for model construction and computation using the saddlepoint approximation. The framework is based on a collection of model-building operations that preserve access to MGFs while allowing complex distributions to be constructed from simpler components. With these components, users need only provide a high-level specification of the model structure, from which the software automatically assembles the necessary generating functions, saddlepoints, and gradients, and performs the optimization of the saddlepoint likelihood. We also introduce a diagnostic that quantifies the difference between saddlepoint and exact likelihood estimates, even when the exact likelihood is unavailable. The framework is implemented in the R package saddlepoint and provides fast, convenient computation of parameter estimates, standard errors, and the discrepancy diagnostic. Numerous examples illustrate the scope and flexibility of the approach.

stat.CO↗

arXiv2026

Potential Estimates and Hodge Systems with $L^1$ data on compact manifolds

In this paper we establish optimal Lorentz estimates for the Riesz potentials acting on closed or co-closed $k$-forms of finite mass on a smooth, compact Riemannian manifold of dimension $n$: For $\alpha \in (0,n)$ and $k=1,\ldots,n-1$, there exists a constant $C>0$ such that \begin{align*} \| \mathcal{I}_{\alpha,k} F \|_{L^{n/(n-\alpha),1}(\Lambda^k)} \leq C \| F\|_{L^1(\Lambda^k)} \end{align*} for all $k$-forms $F \in L^1(\Lambda^k)$ orthogonal to the space of harmonic $k$-forms and satisfying $\mathrm{d} F=0$ or $\mathrm{d}^* F=0$. We show how this inequality implies analogous Lorentz bounds for solutions of the $k$-form Poisson equation and for the Hodge system with data having finite mass. These results include as a special case the div--curl system on the $3$-dimensional torus, where we answer an open question originally posed by J. Bourgain and H. Brezis.

math.AP↗

arXiv2025

Leakage-Resilient Extractors against Number-on-Forehead Protocols

Given a sequence of $N$ independent sources $\mathbf{X}_1,\mathbf{X}_2,\dots,\mathbf{X}_N\sim\{0,1\}^n$, how many of them must be good (i.e., contain some min-entropy) in order to extract a uniformly random string? This question was first raised by Chattopadhyay, Goodman, Goyal and Li (STOC '20), motivated by applications in cryptography, distributed computing, and the unreliable nature of real-world sources of randomness. In their paper, they showed how to construct explicit low-error extractors for just $K \geq N^{1/2}$ good sources of polylogarithmic min-entropy. In a follow-up, Chattopadhyay and Goodman improved the number of good sources required to just $K \geq N^{0.01}$ (FOCS '21). In this paper, we finally achieve $K=3$. Our key ingredient is a near-optimal explicit construction of a new pseudorandom primitive, called a leakage-resilient extractor (LRE) against number-on-forehead (NOF) protocols. Our LRE can be viewed as a significantly more robust version of Li's low-error three-source extractor (FOCS '15), and resolves an open question put forth by Kumar, Meka, and Sahai (FOCS '19) and Chattopadhyay, Goodman, Goyal, Kumar, Li, Meka, and Zuckerman (FOCS '20). Our LRE construction is based on a simple new connection we discover between multiparty communication complexity and non-malleable extractors, which shows that such extractors exhibit strong average-case lower bounds against NOF protocols.

cs.CC↗

arXiv2025

What is the price of approximation? The saddlepoint approximation to a likelihood function

The saddlepoint approximation to the likelihood, and its corresponding maximum likelihood estimate (MLE), offer an alternative estimation method when the true likelihood is intractable or computationally expensive. However, maximizing this approximated likelihood instead of the true likelihood inevitably comes at a price: a discrepancy between the MLE derived from the saddlepoint approximation and the true MLE. In previous studies, the size of this discrepancy has been investigated via simulation, or by engaging with the true likelihood despite its computational difficulties. Here, we introduce an explicit and computable approximation formula for the discrepancy, through which the adequacy of the saddlepoint-based MLE can be directly assessed. We present examples demonstrating the accuracy of this formula in specific cases where the true likelihood can be calculated. Additionally, we present asymptotic results that capture the behaviour of the discrepancy in a suitable limiting framework.

stat.ME↗

arXiv2025

An atomic decomposition of one-dimensional metric currents without boundary

This paper proves an atomic decomposition of the space of $1$-dimensional metric currents without boundary, in which the atoms are specified by closed Lipschitz curves with uniform control on their Morrey norms. Our argument relies on a geometric construction which states that for any $\epsilon>0$ one can express a piecewise-geodesic closed curve as the sum of piecewise-geodesic closed curves whose total length is at most $(1+\epsilon)$ times the original length and whose Morrey norms are each bounded by a universal constant times $\epsilon^{-2}$. In Euclidean space, our results refine the state of the art, providing an approximation of divergence free measures by limits of sums of closed polygonal paths whose total length can be made arbitrarily close to the norm of the approximated measure.

math.FA↗

arXiv2024

Improved Condensers for Chor-Goldreich Sources

One of the earliest models of weak randomness is the Chor-Goldreich (CG) source. A $(t,n,k)$-CG source is a sequence of random variables $X=(X_1,\dots,X_t)\sim(\{0,1\}^n)^t$, where each $X_i$ has min-entropy $k$ conditioned on any fixing of $X_1,\dots,X_{i-1}$. Chor and Goldreich proved that there is no deterministic way to extract randomness from such a source. Nevertheless, Doron, Moshkovitz, Oh, and Zuckerman showed that there is a deterministic way to condense a CG source into a string with small entropy gap. They gave applications of such a condenser to simulating randomized algorithms with small error and to certain cryptographic tasks. They studied the case where the block length $n$ and entropy rate $k/n$ are both constant. We study the much more general setting where the block length can be arbitrarily large, and the entropy rate can be arbitrarily small. We construct the first explicit condenser for CG sources in this setting, and it can be instantiated in a number of different ways. When the entropy rate of the CG source is constant, our condenser requires just a constant number of blocks $t$ to produce an output with entropy rate $0.9$, say. In the low entropy regime, using $t=$ poly$(n)$ blocks, our condenser can achieve output entropy rate $0.9$ even if each block has just $1$ bit of min-entropy. Moreover, these condensers have exponentially small error. Finally, we provide strong existential and impossibility results. For our existential result, we show that a random function is a seedless condenser (with surprisingly strong parameters) for any small family of sources. As a corollary, we get new existential results for seeded condensers and condensers for CG sources. For our impossibility result, we show the latter result is nearly tight, by giving a simple proof that the output of any condenser for CG sources must inherit the entropy gap of (one block of) its input.

cs.CC↗

arXiv2024

The saddlepoint approximation factors over sample paths of recursively compounded processes

This paper presents an identity between the multivariate and univariate saddlepoint approximations applied to sample path probabilities for a certain class of stochastic processes. This class, which we term the recursively compounded processes, includes branching processes and other models featuring sums of a random number of i.i.d. terms; and compound Poisson processes and other L\'evy processes in which the additive parameter is itself chosen randomly. For such processes, $\hat{f}_{X_1,\dotsc,X_N | X_0=x_0}(x_1,\dots,x_N) = \prod_{n=1}^N \hat{f}_{X_n | X_0=x_0,\dots,X_{n-1}=x_{n-1}}(x_n),$ where the left-hand side is a multivariate saddlepoint approximation applied to the random vector $(X_1,\dots,X_N)$ and the right-hand side is a product of univariate saddlepoint approximations applied to the conditional one-step distributions given the past. Two proofs are given. The first proof is analytic, based on a change-of-variables identity linking the functions that arise in the respective saddlepoint approximations. The second proof is probabilistic, based on a representation of the saddlepoint approximation in terms of tilted distributions, changes of measure, and relative entropies.

math.PR↗

arXiv2024

Low-Degree Polynomials Are Good Extractors

We prove that random low-degree polynomials (over $\mathbb{F}_2$) are unbiased, in an extremely general sense. That is, we show that random low-degree polynomials are good randomness extractors for a wide class of distributions. Prior to our work, such results were only known for the small families of (1) uniform sources, (2) affine sources, and (3) local sources. We significantly generalize these results, and prove the following. 1. Low-degree polynomials extract from small families. We show that a random low-degree polynomial is a good low-error extractor for any small family of sources. In particular, we improve the positive result of Alrabiah, Chattopadhyay, Goodman, Li, and Ribeiro (ICALP 2022) for local sources, and give new results for polynomial and variety sources via a single unified approach. 2. Low-degree polynomials extract from sumset sources. We show that a random low-degree polynomial is a good extractor for sumset sources, which are the most general large family of sources (capturing independent sources, interleaved sources, small-space sources, and more). Formally, for any even $d$, we show that a random degree $d$ polynomial is an $\varepsilon$-error extractor for $n$-bit sumset sources with min-entropy $k=O(d(n/\varepsilon^2)^{2/d})$. This is nearly tight in the polynomial error regime. Our results on sumset extractors imply new complexity separations for linear ROBPs, and the tools that go into its proof may be of independent interest. The two main tools we use are a new structural result on sumset-punctured Reed-Muller codes, paired with a novel type of reduction between extractors. Using the new structural result, we obtain new limits on the power of sumset extractors, strengthening and generalizing the impossibility results of Chattopadhyay, Goodman, and Gurumukhani (ITCS 2024).

cs.CC↗

arXiv2023

Extractors for Polynomial Sources over $\mathbb{F}_2$

We explicitly construct the first nontrivial extractors for degree $d \ge 2$ polynomial sources over $\mathbb{F}_2^n$. Our extractor requires min-entropy $k\geq n - \tilde{\Omega}(\sqrt{\log n})$. Previously, no constructions were known, even for min-entropy $k\geq n-1$. A key ingredient in our construction is an input reduction lemma, which allows us to assume that any polynomial source with min-entropy $k$ can be generated by $O(k)$ uniformly random bits. We also provide strong formal evidence that polynomial sources are unusually challenging to extract from, by showing that even our most powerful general purpose extractors cannot handle polynomial sources with min-entropy below $k\geq n-o(n)$. In more detail, we show that sumset extractors cannot even disperse from degree $2$ polynomial sources with min-entropy $k\geq n-O(n/\log\log n)$. In fact, this impossibility result even holds for a more specialized family of sources that we introduce, called polynomial non-oblivious bit-fixing (NOBF) sources. Polynomial NOBF sources are a natural new family of algebraic sources that lie at the intersection of polynomial and variety sources, and thus our impossibility result applies to both of these classical settings. This is especially surprising, since we do have variety extractors that slightly beat this barrier - implying that sumset extractors are not a panacea in the world of seedless extraction.

cs.CC↗

arXiv2022

Low-Degree Polynomials Extract from Local Sources

We continue a line of work on extracting random bits from weak sources that are generated by simple processes. We focus on the model of locally samplable sources, where each bit in the source depends on a small number of (hidden) uniformly random input bits. Also known as local sources, this model was introduced by De and Watson (TOCT 2012) and Viola (SICOMP 2014), and is closely related to sources generated by $\mathsf{AC}^0$ circuits and bounded-width branching programs. In particular, extractors for local sources also work for sources generated by these classical computational models. Despite being introduced a decade ago, little progress has been made on improving the entropy requirement for extracting from local sources. The current best explicit extractors require entropy $n^{1/2}$, and follow via a reduction to affine extractors. To start, we prove a barrier showing that one cannot hope to improve this entropy requirement via a black-box reduction of this form. In particular, new techniques are needed. In our main result, we seek to answer whether low-degree polynomials (over $\mathbb{F}_2$) hold potential for breaking this barrier. We answer this question in the positive, and fully characterize the power of low-degree polynomials as extractors for local sources. More precisely, we show that a random degree $r$ polynomial is a low-error extractor for $n$-bit local sources with min-entropy $\Omega(r(n\log n)^{1/r})$, and we show that this is tight. Our result leverages several new ingredients, which may be of independent interest. Our existential result relies on a new reduction from local sources to a more structured family, known as local non-oblivious bit-fixing sources. To show its tightness, we prove a "local version" of a structural result by Cohen and Tal (RANDOM 2015), which relies on a new "low-weight" Chevalley-Warning theorem.

cs.CC↗

arXiv2020

Two Approximation Results for Divergence Free Measures

In this paper we prove two approximation results for divergence free measures. The first is a form of an assertion of J. Bourgain and H. Brezis concerning the approximation of solenoidal charges in the strict topology: Given $F \in M_b(\mathbb{R}^d;\mathbb{R}^d)$ such that $\operatorname*{div} F=0$ in the sense of distributions, there exist oriented $C^1$ loops $\Gamma_{i,l}$ with associated measures $\mu_{\Gamma_{i,l}}$ such that \[ F= \lim_{l \to \infty} \frac{\|F\|_{M_b(\mathbb{R}^d;\mathbb{R}^d)}}{n_l \cdot l} \sum_{i=1}^{n_l} \mu_{\Gamma_{i,l}} \] weakly-star in the sense of measures and \[ \lim_{l \to \infty} \frac{1}{n_l \cdot l} \sum_{i=1}^{n_l} \|\mu_{\Gamma_{i,l}}\|_{M_b(\mathbb{R}^d;\mathbb{R}^d)} = 1. \] The second, which is an almost immediate consequence of the first, is that smooth compactly supported functions are dense in \[ \left\{ F \in M_b(\mathbb{R}^d;\mathbb{R}^d): \operatorname*{div}F=0 \right\} \] with respect to the strict topology.

math.AP↗

arXiv2020

Improved Extractors for Small-Space Sources

We study the problem of extracting random bits from weak sources that are sampled by algorithms with limited memory. This model of small-space sources was introduced by Kamp, Rao, Vadhan and Zuckerman (STOC'06), and falls into a line of research initiated by Trevisan and Vadhan (FOCS'00) on extracting randomness from weak sources that are sampled by computationally bounded algorithms. Our main results are the following. 1. We obtain near-optimal extractors for small-space sources in the polynomial error regime. For space $s$ sources over $n$ bits, our extractors require just $k\geq s\cdot$polylog$(n)$ entropy. This is an exponential improvement over the previous best result, which required $k\geq s^{1.1}\cdot2^{\log^{0.51} n}$ (Chattopadhyay and Li, STOC'16). 2. We obtain improved extractors for small-space sources in the negligible error regime. For space $s$ sources over $n$ bits, our extractors require entropy $k\geq n^{1/2+\delta}\cdot s^{1/2-\delta}$, whereas the previous best result required $k\geq n^{2/3+\delta}\cdot s^{1/3-\delta}$ (Chattopadhyay, Goodman, Goyal and Li, STOC'20). To obtain our first result, the key ingredient is a new reduction from small-space sources to affine sources, allowing us to simply apply a good affine extractor. To obtain our second result, we must develop some new machinery, since we do not have low-error affine extractors that work for low entropy. Our main tool is a significantly improved extractor for adversarial sources, which is built via a simple framework that makes novel use of a certain kind of leakage-resilient extractors (known as cylinder intersection extractors), by combining them with a general type of extremal designs. Our key ingredient is the first derandomization of these designs, which we obtain using new connections to coding theory and additive combinatorics.

cs.CC↗

arXiv2020

Asymptotic accuracy of the saddlepoint approximation for maximum likelihood estimation

The saddlepoint approximation gives an approximation to the density of a random variable in terms of its moment generating function. When the underlying random variable is itself the sum of $n$ unobserved i.i.d. terms, the basic classical result is that the relative error in the density is of order $1/n$. If instead the approximation is interpreted as a likelihood and maximised as a function of model parameters, the result is an approximation to the maximum likelihood estimate (MLE) that can be much faster to compute than the true MLE. This paper proves the analogous basic result for the approximation error between the saddlepoint MLE and the true MLE: subject to certain explicit identifiability conditions, the error has asymptotic size $O(1/n^2)$ for some parameters, and $O(1/n^{3/2})$ or $O(1/n)$ for others. In all three cases, the approximation errors are asymptotically negligible compared to the inferential uncertainty. The proof is based on a factorisation of the saddlepoint likelihood into an exact and approximate term, along with an analysis of the approximation error in the gradient of the log-likelihood. This factorisation also gives insight into alternatives to the saddlepoint approximation, including a new and simpler saddlepoint approximation, for which we derive analogous error bounds. As a corollary of our results, we also obtain the asymptotic size of the MLE error approximation when the saddlepoint approximation is replaced by the normal approximation.

math.ST↗

arXiv2018

On the Approximability of Time Disjoint Walks

We introduce the combinatorial optimization problem Time Disjoint Walks (TDW), which has applications in collision-free routing of discrete objects (e.g., autonomous vehicles) over a network. This problem takes as input a digraph $G$ with positive integer arc lengths, and $k$ pairs of vertices that each represent a trip demand from a source to a destination. The goal is to find a walk and delay for each demand so that no two trips occupy the same vertex at the same time, and so that a min-max or min-sum objective over the trip durations is realized. We focus here on the min-sum variant of Time Disjoint Walks, although most of our results carry over to the min-max case. We restrict our study to various subclasses of DAGs, and observe that there is a sharp complexity boundary between Time Disjoint Walks on oriented stars and on oriented stars with the central vertex replaced by a path. In particular, we present a poly-time algorithm for min-sum and min-max TDW on the former, but show that min-sum TDW on the latter is NP-hard. Our main hardness result is that for DAGs with max degree $\Delta\leq3$, min-sum Time Disjoint Walks is APX-hard. We present a natural approximation algorithm for the same class, and provide a tight analysis. In particular, we prove that it achieves an approximation ratio of $\Theta(k/\log k)$ on bounded-degree DAGs, and $\Theta(k)$ on DAGs and bounded-degree digraphs.

cs.DS↗

arXiv2017

Some remarks on boundary operators of Bessel extensions

In this paper we study some boundary operators of a class of Bessel-type Littlewood-Paley extensions whose prototype is \[\Delta_x u(x,y) +\frac{1-2s}{y} \frac{\partial u}{\partial y}(x,y)+\frac{\partial^2 u}{\partial y^2}(x,y)=0 \text{ for }x\in\mathbb{R}^d, y>0, \\ u(x,0)=f(x) \text{ for }x\in\mathbb{R}^d. \] In particular, we show that with a logarithmic scaling one can capture the failure of analyticity of these extensions in the limiting cases $s=k \in \mathbb{N}$.

math.AP↗

arXiv2015

Properties of the Affine Invariant Ensemble Sampler in high dimensions

We present theoretical and practical properties of the affine-invariant ensemble sampler Markov chain Monte Carlo method. In high dimensions the affine-invariant ensemble sampler shows unusual and undesirable properties. We demonstrate this with an $n$-dimensional correlated Gaussian toy problem with a known mean and covariance structure, and analyse the burn-in period. The burn-in period seems to be short, however upon closer inspection we discover the mean and the variance of the target distribution do not match the expected, known values. This problem becomes greater as $n$ increases. We therefore conclude that the affine-invariant ensemble sampler should be used with caution in high dimensional problems. We also present some theoretical results explaining this behaviour.

stat.CO↗

arXiv2013

Degree distribution of shortest path trees and bias of network sampling algorithms

In this article, we explicitly derive the limiting degree distribution of the shortest path tree from a single source on various random network models with edge weights. We determine the asymptotics of the degree distribution for large degrees of this tree and compare it to the degree distribution of the original graph. We perform this analysis for the complete graph with edge weights that are powers of exponential random variables (weak disorder in the stochastic mean-field model of distance), as well as on the configuration model with edge-weights drawn according to any continuous distribution. In the latter, the focus is on settings where the degrees obey a power law, and we show that the shortest path tree again obeys a power law with the same degree power-law exponent. We also consider random $r$-regular graphs for large $r$, and show that the degree distribution of the shortest path tree is closely related to the shortest path tree for the stochastic mean-field model of distance. We use our results to shed light on an empirically observed bias in network sampling methods. This is part of a general program initiated in previous works by Bhamidi, van der Hofstad and Hooghiemstra [Ann. Appl. Probab. 20 (2010) 1907-1965], [Combin. Probab. Comput. 20 (2011) 683-707], [Adv. in Appl. Probab. 42 (2010) 706-738] of analyzing the effect of attaching random edge lengths on the geometry of random network models.

math.PR↗

arXiv2013

Extremal geometry of a Brownian porous medium

The path W[0,t] of a Brownian motion on a d-dimensional torus T^d run for time t is a random compact subset of T^d. We study the geometric properties of the complement T^d \ W[0,t] for t large and d >= 3. In particular, we show that the largest regions in this complement have a linear scale phi = [(d log t)/(d-2)kt]^{1/(d-2)}, where k is the capacity of the unit ball. More specifically, we identify the sets E for which T^d \ W[0,t] contains a translate of phi E, and we count the number of disjoint such translates. Furthermore, we derive large deviation principles for the largest inradius of T^d \ W[0,t] for t large and the epsilon-cover time of T^d for epsilon small. Our results, which generalise laws of large numbers proved by Dembo, Peres and Rosen, are based on a large deviation principle for the shape of the component with largest capacity in T^d \ W_rho[0,t], where W_rho[0,t] is the Wiener sausage of radius rho = rho(t), with rho(t) chosen much smaller than phi but not too small. The idea behind this choice is that T^d \ W[0,t] consists of "lakes", whose linear size is of order phi, connected by narrow "channels". We also derive large deviation principles for the principal Dirichlet eigenvalue and for the maximal volume of the components of T^d \ W_rho[0,t] for t large. Our results give a complete picture of the extremal geometry of T^d \ W[0,t] and of the optimal strategy for W[0,t] to realise the extremes.

math.PR↗