En poursuivant votre navigation sur ce site, vous acceptez l'utilisation d'un simple cookie d'identification. Aucune autre exploitation n'est faite de ce cookie. OK

Probabilités et statistiques 535 résultats

Filtrer
Sélectionner : Tous / Aucun
Q
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

On robustness and local differential privacy - ... (Auteur de la conférence) | H

Multi angle

It is of soaring demand to develop statistical analysis tools that are robust against contamination as well as preserving individual data owners' privacy. In spite of the fact that both topics host a rich body of literature, to the best of our knowledge, we are the first to systematically study the connections between the optimality under Huber's contamination model and the local differential privacy (LDP) constraints. We start with a general minimax lower bound result, which disentangles the costs of being robust against Huber's contamination and preserving LDP. We further study four concrete examples: a two-point testing problem, a potentially-diverging mean estimation problem, a nonparametric density estimation problem and a univariate median estimation problem. For each problem, we demonstrate procedures that are optimal in the presence of both contamination and LDP constraints, comment on the connections with the state-of-the-art methods that are only studied under either contamination or privacy constraints, and unveil the connections between robustness and LDP via partially answering whether LDP procedures are robust and whether robust procedures can be efficiently privatised. Overall, our work showcases a promising prospect of joint study for robustness and local differential privacy.
This is joint work with Mengchu Li and Yi Yu.[-]
It is of soaring demand to develop statistical analysis tools that are robust against contamination as well as preserving individual data owners' privacy. In spite of the fact that both topics host a rich body of literature, to the best of our knowledge, we are the first to systematically study the connections between the optimality under Huber's contamination model and the local differential privacy (LDP) constraints. We start with a general ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Algorithmic robust statistics - ... (Auteur de la conférence) | H

Multi angle

The field of Robust Statistics studies the problem of designing estimators that perform well even when the data significantly deviates from the idealized modeling assumptions. The classical statistical theory, going back to the pioneering works by Tukey and Huber in the 1960s, characterizes the information-theoretic limits of robust estimation for a number of statistical tasks. On the other hand, until fairly recently, the computational aspects of this field were poorly understood. Specifically, no scalable robust estimation methods were known in high dimensions, even for the most basic task of mean estimation.
A recent line of work in computer science developed the first computationally efficient robust estimators in high dimensions for a range of learning tasks. This tutorial will provide an overview of these algorithmic developments and discuss some open problems in the area.[-]
The field of Robust Statistics studies the problem of designing estimators that perform well even when the data significantly deviates from the idealized modeling assumptions. The classical statistical theory, going back to the pioneering works by Tukey and Huber in the 1960s, characterizes the information-theoretic limits of robust estimation for a number of statistical tasks. On the other hand, until fairly recently, the computational aspects ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Privacy: definitions, procedures, open problems - ... (Auteur de la conférence) | H

Multi angle

I will provide a broad overview of differential privacy, which provides guarantees that a data analysis protects the privacy of data contributors. The main focus will be on the private computation and release of different statistics, both classical (low-dimensional) and high-dimensional statistics. In addition to giving a high-level program for the development of optimal private estimators, I will likely discuss a few open questions as well.

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Minimax estimation in Efron's two-groups model - ... (Auteur de la conférence) | H

Multi angle

The advent of large scale inference has spurred reexamination of conventional statistical thinking. In a series of highly original articles, Efron showed in some examples that the ensemble of the null distributed test statistics grossly deviated from the theoretical null distribution, and Efron persuasively illustrated the danger in assuming the theoretical null's veracity for downstream inference. Though intimidating in other contexts, the large scale setting is to the statistician's benefit here. There is now potential to estimate, rather than assume, the null distribution.
In a model for n many z-scores with at most k nonnulls, we adopt Efron's suggestion and consider estimation of location and scale parameters for a Gaussian null distribution. Placing no assumptions on the nonnull effects, we consider rate-optimal estimation in the entire regime k < n/2, that is, precisely the regime in which the null parameters are identifiable. The minimax upper bound is obtained by considering estimators based on the empirical characteristic function and the classical kernel mode estimator. Faster rates than those in Huber's contamination model are achievable by exploiting the Gaussian character of the data. As a consequence, it is shown that consistent estimation is indeed possible in the practically relevant regime k ≍ n. In a certain regime, the minimax lower bound involves constructing two marginal distributions whose characteristic functions match on a wide interval containing zero. The construction notably differs from those in the literature by sharply capturing a second-order scaling of n/2 − k in the minimax rate.[-]
The advent of large scale inference has spurred reexamination of conventional statistical thinking. In a series of highly original articles, Efron showed in some examples that the ensemble of the null distributed test statistics grossly deviated from the theoretical null distribution, and Efron persuasively illustrated the danger in assuming the theoretical null's veracity for downstream inference. Though intimidating in other contexts, the ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y
Statistical fairness seeks to ensure an equitable distribution of predictions or algorithmic decisions across different sensitive groups. Among the fairness criteria under consideration, demographic parity is arguably the most conceptually straightforward: it simply requires that the distribution of outcomes is identical across all sensitive groups. In this talk, we explore the relationship between classification and regression problems under this constraint.
We provide several fundamental characterizations of the optimal classification function under the demographic parity constraint. In the awareness framework, analogous to the classical unconstrained classification scenario, we demonstrate that maximizing accuracy under this fairness constraint is equivalent to solving a fair regression problem followed by thresholding at level 1/2. We extend this result to linear-fractional classification measures (e.g., 𝐹-score, AM measure, balanced accuracy, etc.), emphasizing the pivotal role played by regression in this framework. Our findings leverage the recently developed connection between the demographic parity constraint and the multi-marginal optimal transport formulation. Informally, our result shows that the transition between the unconstrained problem and the fair one is achieved by replacing the conditional expectation of the label by the solution of the fair regression problem. Leveraging our analysis, we also demonstrate an equivalence between the awareness and the unawareness setups for two sensitive groups.[-]
Statistical fairness seeks to ensure an equitable distribution of predictions or algorithmic decisions across different sensitive groups. Among the fairness criteria under consideration, demographic parity is arguably the most conceptually straightforward: it simply requires that the distribution of outcomes is identical across all sensitive groups. In this talk, we explore the relationship between classification and regression problems under ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Mean field results in fluid mechanics - Lecture 1 - ... (Auteur de la conférence) | H

Multi angle

Fluid mechanics is rich in mean field results like those of point vortex approximation in 2D. Deviation from the mean field is also a next step of great interest. We illustrate these facts with the example of particle aggregation in a turbulent fluid, a problem of interest for initial rain formation or planet formation in stellar dust disks. The particles have inertia, measured by the so-called Stokes number. When Stokes is large, the mean field theory describes reality well and produces physical laws coherent with experiments. But when Stokes is small, the mean field is not sufficient and a complete solution is still debated. Rigorous elements of the theory and heuristics about the Physics will be given.[-]
Fluid mechanics is rich in mean field results like those of point vortex approximation in 2D. Deviation from the mean field is also a next step of great interest. We illustrate these facts with the example of particle aggregation in a turbulent fluid, a problem of interest for initial rain formation or planet formation in stellar dust disks. The particles have inertia, measured by the so-called Stokes number. When Stokes is large, the mean field ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Numerical methods for SDEs with singular coefficients - Lecture 1 - ... (Auteur de la conférence) | H

Multi angle

We will consider the discretization of the stochastic differential equation$$X_t=X_0+W_t+\int_0^t b\left(s, X_s\right) d s, t \in[0, T]$$where the drift coefficient $b:[0, T] \times \mathbb{R}^d \rightarrow \mathbb{R}^d$ is measurable and satisfies the integrability condition : $\|b\|_{L^q\left([0, T], L^\rho\left(\mathbb{R}^d\right)\right)}<\infty$ for some $\rho, q \in(0,+\infty]$ such that$$\rho \geq 2 \text { and } \frac{d}{\rho}+\frac{2}{q}<1 .$$Krylov and Röckner [3] established strong existence and uniqueness under this condition.Let $n \in \mathbb{N}^*, h=\frac{T}{n}$ and $t_k=k h$ for $k \in \left [ \left [0,n \right ] \right ]$. Since there is no smoothing effect in the time variable, we introduce a sequence $\left(U_k\right)_{k \in \left [ \left [0,n-1 \right ] \right ]}$ independent from $\left(X_0,\left(W_t\right)_{t \geq 0}\right)$ of independent random variables which are respectively distributed according to the uniform law on $[k h,(k+1) h]$. The resulting scheme Euler is initialized by $X_0^h=X_0$ and evolves inductively on the regular time-grid $\left(t_k=k h\right)_{k \in \left [ \left [0,n \right ] \right ]}$ by:$$X_{t_{k+1}}^h=X_{t_k}^h+W_{t_{k+1}}-W_{t_k}+b_h\left(U_k, X_{t_k}^h\right) h$$where $b_h$ is some truncation of the drift function $b$. When $b$ is bounded, one of course chooses $b_h=b$. Then the order of weak convergence in total variation distance is $1 / 2$, as proved in [1]. It improves to 1 up to some logarithmic correction under some additional uniform in time bound on the spatial divergence of the drift coefficient. In the general case (1), we will see that for suitable truncations $b_h$, the difference between the transition densities of the stochastic differential equation and its Euler scheme is bounded from above by $C h^{\frac{1}{2}\left(1-\left(\frac{d}{\rho}+\frac{2}{q}\right)\right)}$ multiplied by some centered Gaussian density, as proved in [2].[-]
We will consider the discretization of the stochastic differential equation$$X_t=X_0+W_t+\int_0^t b\left(s, X_s\right) d s, t \in[0, T]$$where the drift coefficient $b:[0, T] \times \mathbb{R}^d \rightarrow \mathbb{R}^d$ is measurable and satisfies the integrability condition : $\|b\|_{L^q\left([0, T], L^\rho\left(\mathbb{R}^d\right)\right)}<\infty$ for some $\rho, q \in(0,+\infty]$ such that$$\rho \geq 2 \text { and } \frac{d}{\rho}+\f...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Mean field results in fluid mechanics - Lecture 2 - ... (Auteur de la conférence) | H

Multi angle

Fluid mechanics is rich in mean field results like those of point vortex approximation in 2D. Deviation from the mean field is also a next step of great interest. We illustrate these facts with the example of particle aggregation in a turbulent fluid, a problem of interest for initial rain formation or planet formation in stellar dust disks. The particles have inertia, measured by the so-called Stokes number. When Stokes is large, the mean field theory describes reality well and produces physical laws coherent with experiments. But when Stokes is small, the mean field is not sufficient and a complete solution is still debated. Rigorous elements of the theory and heuristics about the Physics will be given.[-]
Fluid mechanics is rich in mean field results like those of point vortex approximation in 2D. Deviation from the mean field is also a next step of great interest. We illustrate these facts with the example of particle aggregation in a turbulent fluid, a problem of interest for initial rain formation or planet formation in stellar dust disks. The particles have inertia, measured by the so-called Stokes number. When Stokes is large, the mean field ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Mean field results in fluid mechanics - Lecture 3 - ... (Auteur de la conférence) | H

Multi angle

Fluid mechanics is rich in mean field results like those of point vortex approximation in 2D. Deviation from the mean field is also a next step of great interest. We illustrate these facts with the example of particle aggregation in a turbulent fluid, a problem of interest for initial rain formation or planet formation in stellar dust disks. The particles have inertia, measured by the so-called Stokes number. When Stokes is large, the mean field theory describes reality well and produces physical laws coherent with experiments. But when Stokes is small, the mean field is not sufficient and a complete solution is still debated. Rigorous elements of the theory and heuristics about the Physics will be given.[-]
Fluid mechanics is rich in mean field results like those of point vortex approximation in 2D. Deviation from the mean field is also a next step of great interest. We illustrate these facts with the example of particle aggregation in a turbulent fluid, a problem of interest for initial rain formation or planet formation in stellar dust disks. The particles have inertia, measured by the so-called Stokes number. When Stokes is large, the mean field ...[+]

Sélection Signaler une erreur
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y
It has been known for a long time that Hamilton-Jacobi-Bellman (HJB) equations preserve convexity, namely if the terminal condition is convex, the solution stays convex at all times. Equivalently, log-concavity is preserved along the heat equation, namely if one starts with a log-concave density, then the solution stays log-concave at all times. Both these facts are a direct consequence of Prékopa-Leindler inequality. In this talk, I will illustrate how a careful second-order analysis on coupling by reflection on the characteristics of the HJB equation reveals the existence of weaker notions of convexity that propagate backward along HJB. More precisely, by introducing the notion of integrated convexity profile, we are able to construct families of functions that fail to be convex, but are still invariant under the action of the HJB equation. In the second part of the talk I will illustrate some applications of these invariance results to the exponential convergence of learning algorithms for entropic optimal transport.[-]
It has been known for a long time that Hamilton-Jacobi-Bellman (HJB) equations preserve convexity, namely if the terminal condition is convex, the solution stays convex at all times. Equivalently, log-concavity is preserved along the heat equation, namely if one starts with a log-concave density, then the solution stays log-concave at all times. Both these facts are a direct consequence of Prékopa-Leindler inequality. In this talk, I will ...[+]

Sélection Signaler une erreur