Searcharxiv⌕ Search

arXiv subjects

Jose H. Blanchet

Publications and source records attributed to Jose H. Blanchet.

10 recordsLinked to original sources
arXiv2026

Optimal Watermark Localization in Mixed-Source Large Language Model Texts

Watermarking provides a principled way to authenticate text generated by large language models (LLMs). In practice, however, the final text may be mixed-source, with watermark evidence surviving at only a subset of token positions after rewriting, insertion, deletion, or paraphrasing. Although prior work has studied global detection of watermark signals, when such signals can be localized remains unclear. We formulate watermark localization as a token-level multiple-testing problem based on pivotal statistics, with a latent indicator recording whether watermark dependence survives at each position. Under an asymptotic regime indexed by exponents for signal sparsity, next-token concentration, and effective-vocabulary growth, we derive a sharp boundary for global detection and phase transitions for discovery and classification within the class of coordinatewise pivot-based localization rules. We show that discovery is strictly harder than detection and that consistent classification is impossible across the parameter regime within this class. We then develop an adaptive thresholding method that does not require knowledge of the exponents or time-varying next-token distributions, but uses a data-driven estimate of the surviving watermark fraction. The method attains the optimal discovery boundary and near-optimal discovery power relative to homogeneous pivot-based rules. Simulations support the theoretical phase transitions, while experiments on model-generated texts demonstrate practical localization performance under common edit mechanisms.

stat.ME↗
arXiv2022

Smoothed Variable Sample-size Accelerated Proximal Methods for Nonsmooth Stochastic Convex Programs

We consider minimizing $f(x) = \mathbb{E}[f(x,ω)]$ when $f(x,ω)$ is possibly nonsmooth and either strongly convex or convex in $x$. (I) Strongly convex. When $f(x,ω)$ is $μ-$strongly convex in $x$, we propose a variable sample-size accelerated proximal scheme (VS-APM) and apply it on $f_η(x)$, the ($η$-)Moreau smoothed variant of $\mathbb{E}[f(x,ω)]$; we term such a scheme as (m-VS-APM). We consider three settings. (a) Bounded domains. In this setting, VS-APM displays linear convergence in inexact gradient steps, each of which requires utilizing an inner (SSG) scheme. Specifically, mVS-APM achieves an optimal oracle complexity in SSG steps; (b) Unbounded domains. In this regime, under a weaker assumption of suitable state-dependent bounds on subgradients, an unaccelerated variant mVS-PM is linearly convergent; (c) Smooth ill-conditioned $f$. When $f$ is $L$-smooth and $κ= L/μ\ggg 1$, we employ mVS-APM where increasingly accurate gradients $\nabla_x f_η(x)$ are obtained by VS-APM. Notably, mVS-APM displays linear convergence and near-optimal complexity in inner proximal evaluations (upto a log factor) compared to VS-APM. But, unlike a direct application of VS-APM, this scheme is characterized by larger steplengths and better empirical behavior; (II) Convex. When $f(x,ω)$ is merely convex but smoothable, by suitable choices of the smoothing, steplength, and batch-size sequences, smoothed VS-APM (or sVS-APM) produces sequences for which expected sub-optimality diminishes at the rate of $\mathcal{O}(1/k)$ with an optimal oracle complexity of $\mathcal{O}(1/ε^2)$. Finally, sVS-APM and VS-APM produce sequences that converge almost surely to a solution of the original problem.

math.OC↗
arXiv2022

Modeling Extremes with d-max-decreasing Neural Networks

We propose a novel neural network architecture that enables non-parametric calibration and generation of multivariate extreme value distributions (MEVs). MEVs arise from Extreme Value Theory (EVT) as the necessary class of models when extrapolating a distributional fit over large spatial and temporal scales based on data observed in intermediate scales. In turn, EVT dictates that $d$-max-decreasing, a stronger form of convexity, is an essential shape constraint in the characterization of MEVs. As far as we know, our proposed architecture provides the first class of non-parametric estimators for MEVs that preserve these essential shape constraints. We show that our architecture approximates the dependence structure encoded by MEVs at parametric rate. Moreover, we present a new method for sampling high-dimensional MEVs using a generative model. We demonstrate our methodology on a wide range of experimental settings, ranging from environmental sciences to financial mathematics and verify that the structural properties of MEVs are retained compared to existing methods.

stat.ML↗
arXiv2021

Asymptotically Optimal Control of a Centralized Dynamic Matching Market with General Utilities

We consider a matching market where buyers and sellers arrive according to independent Poisson processes at the same rate and independently abandon the market if not matched after an exponential amount of time with the same mean. In this centralized market, the utility for the system manager from matching any buyer and any seller is a general random variable. We consider a sequence of systems indexed by $n$ where the arrivals in the $n^{\mathrm{th}}$ system are sped up by a factor of $n$. We analyze two families of one-parameter policies: the population threshold policy immediately matches an arriving agent to its best available mate only if the number of mates in the system is above a threshold, and the utility threshold policy matches an arriving agent to its best available mate only if the corresponding utility is above a threshold. Using a fluid analysis of the two-dimensional Markov process of buyers and sellers, we show that when the matching utility distribution is light-tailed, the population threshold policy with threshold $\frac{n}{\ln n}$ is asymptotically optimal among all policies that make matches only at agent arrival epochs. In the heavy-tailed case, we characterize the optimal threshold level for both policies. We also study the utility threshold policy in an unbalanced matching market with heavy-tailed matching utilities and find that the buyers and sellers have the same asymptotically optimal utility threshold. We derive optimal thresholds when the matching utility distribution is exponential, uniform, Pareto, and correlated Pareto. We find that as the right tail of the matching utility distribution gets heavier, the threshold level of each policy (and hence market thickness) increases, as does the magnitude by which the utility threshold policy outperforms the population threshold policy.

math.PR↗
arXiv2019

Unbiased Multilevel Monte Carlo: Stochastic Optimization, Steady-state Simulation, Quantiles, and Other Applications

We present general principles for the design and analysis of unbiased Monte Carlo estimators in a wide range of settings. Our estimators posses finite work-normalized variance under mild regularity conditions. We apply our estimators to various settings of interest, including unbiased optimization in Sample Average Approximations, unbiased steady-state simulation of regenerative processes, quantile estimation and nested simulation problems.

math.ST↗
arXiv2018

On Optimal Exact Simulation of Max-Stable and Related Random Fields

We consider the random field M(t)=\sup_{n\geq 1}\big\{-\log A_{n}+X_{n}(t)\big\}\,,\qquad t\in T\, for a set $T\subset \mathbb{R}^{m}$, where $(X_{n})$ is an iid sequence of centered Gaussian random fields on $T$ and $0 0$, samples $M(t_{1}),\ldots ,M(t_{d})$ with complexity $o(c(d)\,d^{ε})$. Moreover, if $X_{n}$ has an a.s. converging series representation, then $M$ can be a.s. approximated with error $δ$ uniformly over $T$ and with complexity $O(1/(δ\log (1/δ))^{1/α})$, where $α$ relates to the Hölder continuity exponent of the process $X_{n}$ (so, if $X_{n}$ is Brownian motion, $α=1/2$).

math.PR↗
arXiv2012

Efficient Monte Carlo for high excursions of Gaussian random fields

Our focus is on the design and analysis of efficient Monte Carlo methods for computing tail probabilities for the suprema of Gaussian random fields, along with conditional expectations of functionals of the fields given the existence of excursions above high levels, b. Naïve Monte Carlo takes an exponential, in b, computational cost to estimate these probabilities and conditional expectations for a prescribed relative accuracy. In contrast, our Monte Carlo procedures achieve, at worst, polynomial complexity in b, assuming only that the mean and covariance functions are Hölder continuous. We also explain how to fine tune the construction of our procedures in the presence of additional regularity, such as homogeneity and smoothness, in order to further improve the efficiency.

math.PR↗
arXiv2012

A Practical Implementation of the Bernoulli Factory

The Bernoulli Factory is an algorithm that takes as input a series of i.i.d. Bernoulli random variables with an unknown but fixed success probability $p$, and outputs a corresponding series of Bernoulli random variables with success probability $f(p)$, where the function $f$ is known and defined on the interval $[0,1]$. While several practical uses of the method have been proposed in Monte Carlo applications, these require an implementation framework that is flexible, general and efficient. We present such a framework for functions that are either strictly linear, concave, or convex on the unit interval using a series of envelope functions defined through a cascade, and show that this method not only greatly reduces the number of input bits needed in practice compared to other currently proposed solutions for more specific problems, and is easy to specify for simple forms, but can easily be coupled to asymptotically efficient methods to allow for theoretically strong results.

stat.AP↗
arXiv2009

Efficient importance sampling for binary contingency tables

Importance sampling has been reported to produce algorithms with excellent empirical performance in counting problems. However, the theoretical support for its efficiency in these applications has been very limited. In this paper, we propose a methodology that can be used to design efficient importance sampling algorithms for counting and test their efficiency rigorously. We apply our techniques after transforming the problem into a rare-event simulation problem--thereby connecting complexity analysis of counting problems with efficiency in the context of rare-event simulation. As an illustration of our approach, we consider the problem of counting the number of binary tables with fixed column and row sums, $c_j$'s and $r_i$'s, respectively, and total marginal sums $d=\sum_jc_j$. Assuming that $\max_jc_j=o(d^{1/2})$, $\sum c_j^2=O(d)$ and the $r_j$'s are bounded, we show that a suitable importance sampling algorithm, proposed by Chen et al. [J. Amer. Statist. Assoc. 100 (2005) 109--120], requires $O(d^3\varepsilon^{-2}δ^{-1})$ operations to produce an estimate that has $\varepsilon$-relative error with probability $1-δ$. In addition, if $\max_jc_j=o(d^{1/4-δ_0})$ for some $δ_0>0$, the same coverage can be guaranteed with $O(d^3\varepsilon^{-2}\log(δ^{-1}))$ operations.

math.PR↗