Sara Pérez Vieites Postdoctoral Researcher

About me

I'm a postdoctoral researcher at the School of Electrical Engineering, Aalto University (Espoo, Finland), and I am affiliated with the Finnish Center for Artificial Intelligence (FCAI). I am part of the Cyber-physical Systems group, lead by Dr. Dominik Baumann. Just before starting this position, I was a postdoctoral researcher at IMT Nord Europe (Lille, France) and a visiting researcher at the University of Edinburgh, working mainly with Prof. Víctor Elvira. I finished the Ph.D. in Multimedia and Communications (Statistical Signal Processing) from Universidad Carlos III de Madrid in 2022, under the supervision of Prof. Joaquín Míguez.

My research interests are focused on computational statistics, data assimilation, and signal processing. More specifically, I am interested in Bayesian inference in state-space models as well as in Bayesian experimental design. I work on providing new techniques that run recursively (online) with reduced computational complexity (compared to the state-of-the-art methods) to obtain both parameter and state estimates. I'm also interested in the application of these probabilistic methods in different fields of science such as ecology, energy, geoscience, and climate.


Home

Location: Glenmore Forest Park, Loch Morlich, Aviemore, Scotland.

Teaching and Resources

I have created some informal lecture materials and examples to help understand Monte Carlo methods and related techniques. You can find these materials on my GitHub repository:

Mini Lectures on Monte Carlo Methods

Last news

  • Sep. 24 |   New arXiv preprint! Title: Recursive nested filtering for efficient amortized Bayesian experimental design. [arXiv]
  • June 24 |   Joining both the BAYSM 2024 and 2024 ISBA World meeting next week! I will be giving a talk (29th June) and presenting a poster (5th July) on parameter-space exploration in nested filtering. See you in Venice!
  • Apr. 24 |   Published paper in ICASSP 2024! Title: End-to-End Learning of Gaussian Mixture Proposals Using Differentiable Particle Filters and Neural Networks. [DOI] [pdf]

All news