Lionel Riou-Durand

Research

My research has led to the following outputs. My scholar profile is accessible here.

Preprints

2022 - C.C. Margossian, M.D. Hoffman, P. Sountsov, L. Riou-Durand, A. Vehtari, A. Gelman. Nested R̂: Assessing the convergence of Markov chain Monte Carlo when running many short chains. arXiv:2110.13017 (to appear in Bayesian Analysis).

[paper | code]

2022 - L. Riou-Durand, J. Vogrinc. Metropolis Adjusted Langevin Trajectories: a robust alternative to Hamiltonian Monte Carlo. arXiv:2202.13230.

[paper | poster | slides | 20' video | 50' video]

Publications

2023 - L. Riou-Durand, P. Sountsov, J. Vogrinc, C.C. Margossian, S. Power. Adaptive Tuning for Metropolis Adjusted Langevin Trajectories. Proceedings of the 26th International Conference on Artificial Intelligence and Statistics, PMLR 206:8102-8116.

[paper | code | slides | 15' video]

2022 - A. Dalalyan, A. Karagulyan, L. Riou-Durand. Bounding the error of discretized Langevin algorithms for non-strongly log-concave targets. Journal of Machine Learning Research 23 (235), 1-38.

[paper]

2020 - A. Dalalyan, L. Riou-Durand. On sampling from a log-concave density using kinetic Langevin diffusions. Bernoulli 26 (3), 1956-1988.

[paper | slides]

2018 - L. Riou-Durand, N. Chopin. Noise contrastive estimation: Asymptotic properties, formal comparison with MC-MLE. Electronic Journal of Statistics 12 (2), 3473-3518.

[paper | poster | slides]

Packages

I have developed the R Package:

malt - [R-universe | Github] - Implements the sampling algorithm: Metropolis Adjusted Langevin Trajectories (Riou-Durand and Vogrinc, 2022). The MALT algorithm is a robust extension of Hamiltonian Monte Carlo, for which robustness of tuning is enabled by a positive choice of damping parameter (a.k.a friction). Get started and copy-paste the following lines in R:

options(repos = c(lrioudurand = 'https://lrioudurand.r-universe.dev', CRAN = 'https://cloud.r-project.org'))

install.packages('malt')

help(package = 'malt')

Research activities

I am a former organiser of a seminar/reading group about Piecewise Determnistic Markov Processes (PDMP), at the University of Warwick during 2021-2022. The seminar was organized in weekly sessions to which members of the Statistics Department and external researchers were invited.

I am a former organiser of the weekly PhD students' seminar in ENSAE - CREST during 2016-2017.

I have also been in charge of referee reports for several scientific journals and conferences, among which: AoS, JRSSB, JASA, JMLR, Biometrika, JCGS, EJS, COLT.

Talks

MCM 2023, Sorbonne Université, Paris (France), 29 June 2023 (in person).
BayesComp 2023, Levi (Finland), 15 March 2023 (in person).
University College London, DeepMind CSML Seminar Series, London (UK), 7 December 2022 (in person).
CIRM, Fusion Workshop, Luminy (France), 27 October 2022 (in person).
University of Warwick, Statistics 50th Anniversary, Coventry (UK), 9 September 2022 (in person).
MCQMC 2022, Special Sessions, Linz (Austria), 19 July 2022 (in person).
University of Bonn, Oberseminar Stochastics, Bonn (Germany), 7 July 2022 (in person).
ISBA 2022, Invited Sessions, Montréal (Canada), 30 June 2022 (in person).
Université de Montréal, BAYSM 2022, Montréal (Canada), 23 June 2022 (in person).
University of Cambridge, CCIMI Seminar Series, Cambridge (UK), 11 May 2022 (in person).
National University of Singapore, Statistics Seminar Series, Singapore, 11 May 2022 (online).
University of British Columbia, Statistics Seminar Series, Vancouver (Canada), 3 May 2022 (online).
Boston University, Statistics and Probability Seminar Series, Boston (USA), 14 April 2022 (in person).
Harvard University, Bayesian working group, Boston (USA), 13 April 2022 (online).
Columbia University, Statistics Seminar Series, New-York (USA), 11 April 2022 (in person).
New-York University, Complexity of Sampling Seminar, New-York (USA), 8 April 2022 (in person).
Flatiron Institute, Bayesian Seminar, New-York (USA), 8 April 2022 (in person).
INRIA Paris, Conférence ANR QuAMProcs, Paris (France), 9 March 2022 (in person).
Institut Henri Poincaré, Séminaire Parisien de Statistique, Paris (France), 7 March 2022 (in person).
University of Oxford, CSML Seminar Series, Oxford (UK), 18 February 2022 (in person).
University of Florida, Statistics Seminar Series, Gainesville (USA), 17 February 2022 (online).

Lancaster University, CSML Seminar Series, Oxford (UK), 17 February 2022 (online).
RIKEN AIP, Approximate Bayesian Inference Seminar, Tokyo (Japan), 17 February 2022 (online).

Imperial College London, Statistics Seminars (UK), 11 February 2022 (online).
University of Bristol, Statistics Seminars (UK), 4 February 2022 (online).
University of Warwick, YRM Seminars, Coventry (UK), 25 January 2022 (in person).
Université de Nantes, Séminaire de Maths Appliquées, Nantes (France), 11 January 2022 (in person).
ENSAI, Séminaire de Statistique de l'IRMAR, Bruz (France), 7 January 2022 (in person).
CIRM, Meetings in Mathemematical Statistics, Luminy (France), 17 December 2021 (in person).
University of Warwick, ACII Seminars, Coventry (UK), 10 December 2021 (in person).
Alan Turing Institute, DCE Seminar Series, Londres (UK), 16 September 2020 (online).
University of Warwick, ACII Seminars, Coventry (UK), 25 October 2019 (in person).
University of Warwick, YRM Seminars, Coventry (UK), 22 October 2019 (in person).
Sorbonne Université, YRM MathStatistics, Paris (France), 25 September 2018 (in person).
ENSAE, Séminaire de Stats/ML/Econométrie, Palaiseau (France), October 2016 (in person).

Posters

Workshop on Monte Carlo methods, University of Warsaw (Poland), 14 December 2023.
ISBA 2022, Montréal (Canada), 30 July 2022.
Bayes Comp 2020, University of Florida, Gainesville (USA), 7-10 January 2020.
Structural Inference in High-Dimensional Models 2, HSE, St. Petersburg (Russia), 26-30 August 2019.

Google Sites

Report abuse