Javier Selva Castelló
Machine Learning Engineer and Computer Vision Researcher
I received my Bachelor's degree in Computer Science at UPV in 2015. Two years later I got my MSc in Artificial Intelligence from UB, UPC and URV. I recently completed my PhD on Learning Video representations at HuPBA lab. under Prof. Sergio Escalera's supervision. My main research interests include Machine Learning, Natural Language Processing and Computer Vision. I am especially interested in learning algorithms, understanding and analyzing their perks for a more informed development of new techniques. I find self-supervised, multi-modal and generative approaches particularly interesting.
Bio and Interests (Resume)
From a very early age I wanted to be an inventor. As I grew older I gradually narrowed it down to AI. So I set out to study Computer Science and got my bachelor degree at Universitat Politècnica de València (UPV - 2015), where I started working on NLP, performing sentiment analysis on Twitter content. From there I came to Barcelona, where I’m currently based, to study a Master on Artificial Intelligence. Universitat Politècnica de Catalunya (UPC), Universitat de Barcelona (UB) and Universitat Rovila i Virgili (URV), offered a wide program, ranging from traditional symbolic AI to deep learning, going through multi-agent systems. This gave me an understanding of a plethora of AI techniques and finally allowed me to better define and understand what I was interested in: the learning process. This got me interested into the (neuro)science of human learning. Such interest peaked during my master's thesis: a survey on deep video frame prediction (2018), a task which follows predictive coding ideas to train neural networks for computer vision. After that I jumped on a PhD under Prof. Sergio Escalera’s supervision, within the HuPBA lab. My thesis revolved around learning video representations. In particular, I employed the novel transformer architecture, focusing on analyzing human behavior in interactions. This work required a multidisciplinary approach, where I collaborated closely with psychologists, integrating their insights into our projects to bridge the gap between technology and behavioral science. While I’ve found that human analysis is not my preferred line of research, I have strongly enjoyed working with researchers from other fields.
I want to better understand the learning process (what and how are these models learning) in order to design novel mechanisms that allow machines to aquire knowledge, as well as harnessing said knowledge to solve real world problems. My (fundamental) research interests lie mostly within self-supervised methods (including generative ones), reinforcement learning, multi-modal approaches (CV + NLP), and interpretable models, with a particular focus on understanding and improving the learning mechanisms of AI. I think we should be working towards more computationally efficient architectures which are better capable of generalizing while requiring less data. This in turn, I think, is a step towards democratizing AI, which currently is heavily dependent on data, which in turn is not so much accessible by everyone. Regarding applied research, I am deeply motivated to collaborate with experts from diverse fields, combining their knowledge with AI to develop innovative solutions to real-world challenges. I am particularly interested in applying AI to areas such as sustainable agriculture and farming, autonomous driving (for public transport and freight), logistics optimization, clean energy production (such as novel windmills or solar panels), aiding scientific exploration, etc. I believe that we have an incredible tool at our disposal, and despite the current hype around AI, its true potential lies in thoughtfully designed applications. By collaborating with experts across domains, I hope we can harness AI's power to truly improve our quality of life while minimizing our impact the Earth and its resources.
I want to better understand the learning process (what and how are these models learning) in order to design novel mechanisms that allow machines to aquire knowledge, as well as harnessing said knowledge to solve real world problems. My (fundamental) research interests lie mostly within self-supervised methods (including generative ones), reinforcement learning, multi-modal approaches (CV + NLP), and interpretable models, with a particular focus on understanding and improving the learning mechanisms of AI. I think we should be working towards more computationally efficient architectures which are better capable of generalizing while requiring less data. This in turn, I think, is a step towards democratizing AI, which currently is heavily dependent on data, which in turn is not so much accessible by everyone. Regarding applied research, I am deeply motivated to collaborate with experts from diverse fields, combining their knowledge with AI to develop innovative solutions to real-world challenges. I am particularly interested in applying AI to areas such as sustainable agriculture and farming, autonomous driving (for public transport and freight), logistics optimization, clean energy production (such as novel windmills or solar panels), aiding scientific exploration, etc. I believe that we have an incredible tool at our disposal, and despite the current hype around AI, its true potential lies in thoughtfully designed applications. By collaborating with experts across domains, I hope we can harness AI's power to truly improve our quality of life while minimizing our impact the Earth and its resources.