on 11-11-2020
I recently joined DeepMind as a Research Scientist, to study richer forms of representation that can guide planning and prediction of learning agents into more informed goals. (see more)
I started my research path at IST as Physics Engineering undergrad. Later I enrolled into a dual CMU/Portugal PhD program in Robotics/Machine learning. During this period I studied how spectral methods impacted prediction and planning algorithms. I was affiliated with both the IT and the Institute for Systems and Robotics (ISR) at IST together with the Robotics Institute at Carnegie Mellon University. I had the pleasure to be co-advised by André Martins also from IT, and Geoffrey Gordon and Siddartha Srinivasa from CMU.
After finishing my PhD, I joined Sacoor Brothers as a Data Scientist and later joined Priberam Labs as a Leading Research Scientist, where I was able to participate in different research projects in the area of Natural Language Processing, some of which were connected with IT such as the GoLocal project.
As a student, I participated several times as a mentor and later co-organised, the Lisbon Machine Learning Summer School in close connection with IT.
Currently, I am part of the Reinforcement Learning team at DeepMind and I am affiliated with ISR as an invited researcher. I am currently focused on studying questions related with rich representations for planning and how these can impact useful and informative forms of exploration.
I hope to leave my home in the near future and to continue on building good research collaborations with IT.
on 11-11-2020
After my PhD in 2014, I was offered a research position in Equipment Academy in Beijing, China. (see more)
I completed my PhD research at Instituto de Telecomunicações from November 2011 to October 2013, and was supervised by Professor José M. Bioucas-Dias. I spent the two years studying the topic of interferometric phase estimation via sparse coding in the complex domain. During this period, I addressed interferometric phase image estimation problem. The new sparse-coding-based approach to interferometric phase estimation, termed the SpInPHASE, was proposed with the help of Professor Bioucas. I also attended two courses in IT, “Inverse Problems in Imaging” by Professor Bioucas and “Nonlinear Optimization” by Professor João Xavier.
With the supervision of Bioucas, IT provided me the chance to learn much more beyond my research. I also got the meticulous scholarship from Bioucas and the conditions to enter the research field of interferometric phase image estimation during my studies.
After my PhD in 2014, I was offered a research position in Equipment Academy in Beijing, China. I continued my research in phase denoising by the complex valued dictionary learned from the group sparsity vectors and the research of training multiresolution complex valued dictionaries. Actually, all of my research now is based on my gains in IT. No words can express my grateful thanks to IT.