Doctoral Thesis Defense of Ignacio Peis Aznarte
Ignacio Peis Aznarte, a PhD student in the Signal Processing Group of the University Carlos III de Madrid is defending his doctoral thesis titled «Advanced Inference and Representation Learning Methods in Variational Autoencoders.” on September, 22nd:
- Title: Advanced Inference and Representation Learning Methods in Variational Autoencoders.
- Author: Ignacio Peis Aznarte.
- Supervisors: Antonio Artés Rodríguez, Pablo M. Olmos.
- Event Date: Friday, September 22, 10:30
- Location: Sala de Juntas de Biblioteca. 3º planta. edificio Rey Pastor. Universidad Carlos III de Madrid.
- Short abstract: The work presented in this doctoral thesis focuses on the development of novel methods for improving the state-of-the-art in VAEs. Specifically, three fundamental challenges are addressed: achieving meaningful global latent representations, obtaining highly-flexible priors for learning more expressive models, and improving current approximate inference methods. As a first main contribution, an innovative technique named UG-VAE from Unsupervised-Global VAE, aims to enhance the ability of VAEs in capturing factors of variations at data (local) and group (global) level. By carefully desigining the encoder and the decoder, and throughout conductive experiments, it is demonstrated that UG-VAE is effective in capturing unsupervised global factors from images. Second, a non-trivial combination of highly-expressive Hierarchical VAEs with robust Markov Chain Monte Carlo inference (specifically Hamiltonian Monte Carlo), for which important issues are successfully resolved, is presented. The resulting model, referred to as the Hierarchical Hamiltonian VAE model for Mixed-type incomplete data (HH-VAEM), addresses the challenges associated with imputing and acquiring heterogeneous missing data. Throughout extensive experiments, it is demonstrated that HH-VAEM outperforms existing one-layered and Gaussian baselines in the tasks of missing data imputation and supervised learning with missing features, thanks to its improved inference and expressivity. Furthermore, another relevant contribution is presented, namely a sampling-based approach for efficiently computing the information gain when missing features are to be acquired with HH-VAEM. This approach leverages the advantages of HH-VAEM and is demonstrated to be effective in the same tasks.
Doctoral Thesis Defense of Daniel Barrejon Moreno
Daniel Barrejon Moreno, a PhD student in the Signal Processing Group of the University Carlos III de Madrid is defending his doctoral thesis titled «How can humans leverage machine learning? From Medical Data Wrangling to Learning to Defer to Multiple Experts” on October, 4th:
- Title: How can humans leverage machine learning? From Medical Data Wrangling to Learning to Defer to Multiple Experts.
- Author: Daniel Barrejon Moreno.
- Supervisors: Pablo M. Olmos, Antonio Artés Rodríguez.
- Event Date: Wednesday, October 4, 11:00
- Location: Sala de video 3.S1.08, Universidad Carlos III de Madrid.
- Short abstract: This doctoral thesis has investigated two cases where humans can leverage the power of machine learning: first, as a tool to assist in data wrangling and data understanding problems and second, as a collaborative tool where decision-making can be automated by the machine or delegated to human experts, fostering more transparent and trustworthy solutions.
Article accepted for publication in Internet Interventions
The article «Automatic patient functionality assessment from multimodal data using deep learning techniques – Development and feasibility evaluation» by Emese Sukei, Santiago de Leon Martinez, Pablo M. Olmos and Antonio Artés-Rodríguez has been published in Internet Interventions.
Project granted by Instituto de Salud Carlos III (ISCIII)
PMP22/00084 «Medicina Personalizada (MedPer) en la detección precoz del deterioro cognitivo (DC) preclinico. Desarrollo de un modelo predictivo de riesgo» has been granted to Pablo Martínez Olmos in collaboration with Instituto de Investigación Biomédica de Salamanca (IBSAL), by Instituto de Salud Carlos III within the call «Proyectos de Investigación de Medicina Personalizada de Precisión».
Article accepted for publication in Journal of Medical Internet Research (JMIR)
The article «One–week suicide risk prediction using real-time smartphone monitoring» by Maria Luisa Barrigon, Lorena Romero-Medrano, Pablo Moreno-Muñoz, Alejandro Porras-Segovia, Jorge Lopez-Castroman, Philippe Courtet, Antonio Artés-Rodríguez and Enrique Baca-Garcia has been published in JMIR.
Project granted by Instituto de Salud Carlos III (ISCIII)
PMP22/00032 « Integrating longitudinal patient-generated data and multi-omic profiling for comprehensive precision oncology in womens’ cancers» has been granted to Antonio Artés Rodríguez in collaboration with Centro Nacional de Investigaciones Oncológicas (CNIO) by Instituto de Salud Carlos III within the call «Proyectos de Investigación de Medicina Personalizada de Precisión».
Doctoral Thesis Defense of Fernando Moreno Pino
Fernando Moreno Pino, a PhD student in the Signal Processing Group of the University Carlos III de Madrid is defending his doctoral thesis titled «Deep attentive time series modelling for quantitative finance”” on May, 5th:
- Title: Deep Attentive Time Series Modelling for Quantitative Finance.
- Author: Fernando Moreno Pino.
- Supervisors: Antonio Artés Rodríguez, Pablo M. Olmos.
- Event Date: Friday, May 5, 11:00
- Location: Aula de Grados del Padre Soler (Auditorium building)
- Short abstract: Time series modelling and forecasting is a persistent problem with extensive implications in scientific, business, industrial, and economic areas. This thesis’ contribution is twofold. Firstly, we propose a novel probabilistic time series forecasting methodology that introduces the use of Fourier domain-based attention models. We denote Spectral Attention (SA) to this frequency domain-based attention mechanism, which merges classic signal processing spectral filtering techniques with machine learning architectures. Secondly, in the context of high-frequency trading, we take advantage of the abundance of intraday financial data to develop deep learning-based solutions for modelling financial time series.
Article accepted for publication in Neural Networks
The article «Regularizing transformers with deep probabilistic layers” by Aurora Cobo Aguilera, Pablo M. Olmos, Antonio Artés-Rodríguez and Fernando Pérez-Cruz has been published in Neural Networks.
«Living Lab for Assistive Technologies and Artificial Intelligence» Opening Day
The opening of the «Living Lab for Assistive Technologies and Artificial Intelligence» is close, this new infrastructure will house the complete cycle of development of technologies for diagnostic aid, monitoring and human-machine interaction for health care.
Date: 23rd of february at 13:00
Place: Parque Científico UC3M,
If you are interested in attending, you can register for the event at the following link.
Póster – “Inteligencia Artificial para la medición de PROs en pacientes Oncológicos”
In the «1ª edición del concurso de posters de Farmaimpulso OncoHematología 2023», researchers from the Carlos III University of Madrid in collaboration with the Gregorio Marañon Hospital have won 1st Prize in the «Innovative Idea/Project» category. This is a national call that aims to share real-life clinical experiences within the field of oncohematology.
Doctoral Thesis Defense of Lorena Romero Medrano
Lorena Romero Medrano, a PhD student in the Signal Processing Group of the University Carlos III de Madrid is defending her doctoral thesis titled “Change-Point Detection Methods for Behavioral Shift Recognition in Mental Healthcare” on January, 20th:
- Title: Change-Point Detection Methods for Behavioral Shift Recognition in Mental Healthcare
- Author: Lorena Romero (GTS)
- Supervisors: Pablo M. Olmos, Antonio Artés
- Event Date: Friday, January 20, 12:00 – 13:30
- Location: Salón de Grados (Auditorium building)
- Short abstract: The motivation of this thesis is based on the improvement of psychiatric patients assessment. Specially, in the suicide prevention problem, where we work under the hypothesis that behavioral changes in the digital phenotype of a patient might precede new relapses, that we aim to predict in advance. From a technical perspective, we focus on the development of probabilistic models for change-point detection for heterogeneous, high-dimensional and incomplete observations that characterize the previously mentioned clinical scenario.
«Passive Radar» project granted by Office of Naval Research (ONR)
«Advanced detection algorithms for passive radar (Passive Radar)» has been granted to David Ramírez García by the Office of Naval Research (ONR) within the call «ONR BAA Announcement #N00014-22-S-B001».