Hugo Queniat

Hi there! I am a second-year PhD student in the Statistics CDT at the University of Warwick, working under the joint supervision of Prof. Gareth Roberts and Dr. Nick Tawn. Prior to this, I graduated with an MRes. in Mathematics, Vision and Learning ("MVA") at École Normale Supérieure Paris-Saclay and an MSc. in Engineering from Télécom Paris.

Research interests: I am broadly interested in the study of problems at the interface of learning and sampling. Currently, I am working on tempering techniques to sample from multimodal target distributions. My work is supported by the OCEAN ERC Synergy grant.

news

Aug 01, 2025	After a successful research retreat in Paris, I will co-organize the Acceleration Week in Warwick in September for the OCEAN ERC project, bringing together researchers to discuss advancements in sampling methods. Excited for a week of collaboration!

selected publications

arXiv

Annealed Leap-Point Sampler for Multimodal Target Distributions

Nicholas G. Tawn, Matthew T. Moores, Hugo Queniat, and Gareth O. Roberts

2025

Abs PDF

In Bayesian statistics, exploring high-dimensional multimodal posterior distributions poses major challenges for existing MCMC approaches. This paper introduces the Annealed Leap-Point Sampler (ALPS), which augments the target distribution state space with modified annealed (cooled) distributions, in contrast to traditional tempering approaches. The coldest state is chosen such that its annealed density is well-approximated locally by a Laplace approximation. This allows for automated setup of a scalable mode-leaping independence sampler. ALPS requires an exploration component to search for the mode locations, which can either be run adaptively in parallel to improve these mode-jumping proposals, or else as a pre-computation step. A theoretical analysis shows that for a d-dimensional problem the coolest temperature level required only needs to be linear in dimension, {}mathcal{O}(d) implying that the number of iterations needed for ALPS to converge is {}mathcal{O}(d) (typically leading to overall complexity {}mathcal{O}(d^3) when computational cost per iteration is taken into account). ALPS is illustrated on several complex, multimodal distributions that arise from real-world applications. This includes a seemingly-unrelated regression (SUR) model of longitudinal data from U.S. manufacturing firms, as well as a spectral density model that is used in analytical chemistry for identification of molecular biomarkers.