Paul Rognon-Vael
About me:
I am PhD student in statistics at Universitat Pompeu Fabra (UPF) and Universitat Politècnica de Catalunya (UPC). My advisors are David Rossell and Piotr Zwiernik .
Prior to joining UPF, I was a research intern at Marta Melé's lab in the Barcelona Supercomputing Center, working on transcriptomics. In past lives, I was a quantitative finance analyst working on all kinds of risks (climate, market, credit, liquidity, operational, fair lending) for large and midsize banks in the US.
Research interets:
My current work focuses on high-dimensional problems at crossroads between Bayesian and frequentist theory. I am working on designing better regularisation for settings where external / prior information is available on the parameter vector sparsity or the variable selection problem. This kind of information arises naturally from domain expert knowledge, in data integration settings and sequential problems. I developed theoretical results high-dimensional linear regression showing that regularizers informed of block structure in sparsity outperform their non-informed counterparts. I plan on extending this work to other settings such as causal discovery, reinforcement learning or large covariance matrix learning.
From my professional and academic experience, I also developed interests in: graphs and networks, learning under fairness constraints, extreme values, spatial processes and functional data. My favourite areas of applications are genetics and finance / risk.
News:
So I will go to my first ISBA this 2024 in Venice 🛶! I will be talking at an invited session on Theory and methods for data integration with real and very good Bayesians. I'll my put on my full Bayesian mask 🎭 and rock the carnival.
Yayyy, it's going to be a very busy December (both 😃 and 😭). I received funding to attend the Workshop “Bayesian Statistics and Statistical Learning: New Directions in Algebraic Statistics” organised at IMSI (Chicago) this December 2023. I have been working on my algebra, I am ready 😈.
Aquesta és pels locals. Presentaré al Congrès XL de la SEIO a Elx aquest novembre 2023. Será segurament el mateix tema que a Lisboa (aquí abaix ⬇️): informació externa per a selecció de variables amb més concentració en la integració de dades aquest cop.
I won an student award 😳 to present at the upcoming ICSDS in Lisbon this December 2023. I will be talking about my work on external information and blockwise regularisation for sparsity recovery.
=> You might be interested if you like high-dimensional problems, L0 regularization / information criteria, variable selection in sequential or data integration settings, or stuff that works both the Bayesian and the frequentist ways.I was in Chicago this September ✈️ to attend the workshop "Invitation to Algebraic Statistics and Applications" at IMSI. It was hard but I discovered an entire new stream of fascinating theory and a beautiful community of algebraists 👫👭. Can't wait to buff up my algebra skills.
Publication(s):
García-Pérez R., Ramirez J.M., …, Rognon P. J. , …, Melé M. (2023), “The landscape of expression and alternative splicing variation across human traits”, Cell Genomics, 3(1), 100244. (link)
Talks:
ISBA 2024 - invited session: Theory and methods for data integration (Jun. 2024)
ICSDS 2023 - Student travel award session: External information and blockwise regularisation for sparsity recovery (Dec. 2023)
SEIO 2023 Congress - invited session: Statistical inference with external information: high-dimensional data integration (Nov . 2023)
Teaching:
Lab seminars of the Statistical Inference and Modelling course of the Master's in Data Science, Barcelona School of Economics (BSE), 2021 and 2023
Introductory course in probability, statistics and algebra of the Master's in Data Science, Barcelona School of Economics (BSE), 2021 - 2023 (material to be published here soon)
Bayesian high-dimensional regression at 2023 BSE Summer School
Introduction to Python and R programming, Barcelona School of Economics (BSE), 2022 - 2023
Contact:
Department of Economy and Business, Universitat Pompeu Fabra
Ramon Trias Fargas, 25-27, 08005 Barcelona
Office: 23.101
Email: firstname.rognon @ gmail . com