Thursday at 2pm, organised by
Olivier Fercoq (ML), Anne Sabourin, François Portier (PS)
Coming talks :
(PS&ML) October 18, 2pm : Aurelien Bellet (INRIA)
Title: A Massively Distributed Algorithm for Private Averaging with Malicious Adversaries
Abstract: The amount of personal data collected in our everyday interactions with connected devices offers great opportunities for innovative services fueled by machine learning, as well as raises serious concerns for the privacy of individuals. In this paper, we propose a massively distributed protocol for a large set of users to privately compute averages over their joint data, which can then be used to learn predictive models. Our protocol can find a solution of arbitrary accuracy, does not rely on a third party and preserves the privacy of users throughout the execution in both the honest-but-curious and malicious adversary models. Specifically, we prove that the information observed by the adversary (the set of maliciours users) does not significantly reduce the uncertainty in its prediction of private values compared to its prior belief. The level of privacy protection depends on a quantity related to the Laplacian matrix of the network graph and generally improves with the size of the graph. Furthermore, we design a verification procedure which offers protection against malicious users joining the service with the goal of manipulating the outcome of the algorithm.
(PS&ML) November 8, 2pm : Gaelle Chagny (Université de Rouen)
(PS&ML) November 15, 2pm : Stéphane Canu (LITIS, INSA Rouen)
Past talks :
(PS&ML) October 4, 2pm, C49 : Francis Bach (INRIA)
Gossip of Statistical Observations using Orthogonal Polynomials
(PS&ML) April 12 : Joon Kwon (Ecole Polytechnique)
Stratégies de descente miroir pour la minimisation du regret et l’approchabilité
(ML&PS) May 3: Erwan Scornet (Polytechnique, CMAP)
Consistency and minimax rates of random forests
(ML&PS) April 5: Matthieu Lerasle (CNRS, Psud)
MOM pour l’apprentissage robuste
(ML&PS) January 25 : Arthur Charpentier (Rennes 1)
Insurance: Risk Pooling and Price Segmentation. Using Information in a `Big Data’ Context
(ML&PS) Decembre 14 : Arnaud Guyader (LSTA)
Fleming-Viot particle systems: asymptotic behavior and illustration in molecular dynamics
and Romain Brault (L2S, CentraleSupelec).
Scalable learning of Vector-Valued Functions.
(ML&PS) November 23 : Bernard Delyon (Rennes 1 – IRMAR)
Processus gaussiens changés de temps.
and Emmanuel Soubies (EPFL)
Exact continuous relaxations for the l0-regularized least-squares criteria
(PS) November 9 : Romain Azais (INRIA Nancy)
Inference for conditioned Galton-Watson trees from their Harris path
and Aymeric Dieuleveut (Sierra team)
Bridging the Gap between Constant Step Size Stochastic Gradient Descent and Markov Chains
(ML&PS) October 28 : Rémi Gribonval (INRIA / Panama Team)
Compressive Statistical Learning with Random Feature Moments
and Vincent Duval (INRIA / Mokaplanteam)
A gridless method for super-resolution microscopy
(ML&PS) Septembre 21: Stéphane Robin (Agro ParisTech)
Detecting change-points in the structure of a network: Exact Bayesian inference
(ML&PS) Septembre 14: Johan Segers (Université catholique de Louvain)
Accelerating the convergence rate of Monte Carlo integration through ordinary least squares
and Randal Douc (Télécom SudParis)
Posterior consistency for partially observed Markov models
(ML&PS) April 13 : Rémi Bardenet (CNRS and University of Lille)
Monte Carlo with determinantal point processes
and Hong-Phuong Dang (University of Lille)
Bayesian nonparametric approaches and dictionary learning for inverse problems in image processing
(ML&PS) April 6: Gwennaëlle Mabon
Aggregation of Laguerre density estimators
and Jérémie Sublime
Collaborative Clustering and its Applications
March 23:
Julie Josse (polytechnique),
Low-rank interaction log-linear model for contingency table analysis
Balamurugan Palaniappan (LTCI)
Stochastic Variance Reduction Methods for Saddle-point Optimization Problems
March 16:
François Roueff (LTCI)
Introduction aux séries temporelles localement stationaires
Tobias Kley (LSE)
Predictive, finite-sample model choice for time series under stationarity and non-stationarity
March 9:
Jean David Fermanian (ENSAE)
Conditional copulas and some tests of the “simplifying assumption
Olivier Lopez (LSTA, Paris 6)
Copula estimation under censoring and applications in actuarial sciences
L’équipe est impliquée dans l’organisation de plusieurs séminaires à Paris :
- Statistical Machine Learning in Paris (bi-hebdomadaire, lundi et jeudi après-midi, corresp. STA Joseph Salmon)
- Séminaire Parisien de Statistiques (mensuel, lundi après-midi, corresp. STA Joseph Salmon)
- Séminaire BIG’MC (Méthodes de Monte Carlo) (mensuel, jeudi après-midi, corresp. STA Gersende Fort) – fin en Juin 2014.
The team is involved in the organization of several seminars taking place in Paris :
- Statistical Machine Learning in Paris (twice every month, Monday or Thursday afternoon, corresp. STA Olivier Cappé)
- Séminaire parisien de statistiques (every month, Monday afternoon, corresp. STA Joseph Salmon)
- Séminaire BIG’MC (Monte Carlo methods) (evry month, Thursday afternoon, corresp. STA Gersende Fort)
The team seminars are announced in the news section of the site.