Home

Welcome to my web page. Here you will find information about my research activities (for recent works please check publications and blog).

My research interests include machine learning (mostly deep learning) applied to computer vision and multimedia, and artificial intelligence in general. I have worked on topics such as generative models (GANs in particular), transfer learning, continual learning, multimodal representations and neural image compression.

In a broader sense, I am increasingly interested in investigating artificial intelligence within its social and environmental context, and their effects and interplays.

Latest news

More news...

Latest posts

Can we perform unsupervised domain adaptation without accessing source data? Recent works show that it is not only possible but also very effective. In this post I review our recent works (ICCV 2021, NeurIPS 2021, CVIU 2023 and TPAMI 2023), ...
Unmanned vehicles require large amounts of diverse data to train their machine vision modules. Importantly, data should include rare yet important events that the vehicle may face while in autonomous operation. In addition, modern vehicles capture data from multiple cameras ...
Neural image and video codecs achieve competitive rate-distortion performance. However, they have a series of practical limitations, such as relying on heavy models, that hinder their adoption in practice. In this aspect, traditional codecs are usually designed with such practical ...
Neural image codecs typically use specific elements in their architectures, such as GDN layers, hyperpriors and autoregressive context models. These elements allow exploiting contextual redundancy while obtaining accurate estimations of the probability distribution of the bits in the bitstream. Thus, ...
/ deep learning, image compression
Neural image compression (a.k.a. learned image compression) is a new paradigm where codecs are modeled as deep neural networks whose parameters are learned from data. There has been increasing interest in this paradigm as a possible competitor to traditional image ...
/ deep learning, image compression
This is a brief update on mix and match networks (M&MNets), describing the new ideas included in the extended version (IJCV 2020). An earlier post contains more details about the original CVPR 2018 version. Mix and match networks (summary) We ...
The problem of catastrophic forgetting (a network forget previous tasks when learning a new one) and how to address it has been studied mostly in discriminative models such as image classification. In our recent NeurIPS 2018 paper (video), we study ...
Depth sensors capture information that complements conventional RGB data. How to combine them in an effective multimodal representation is still actively studied, and depends on different factors. Here I will focus on scenes and discuss several approaches to RGB-D scene ...
/ deep learning, RGB-D, transfer learning
We recently explored how we can take multiple seen image-to-image translators and reuse them to infer other unseen translations, in an approach we call mix and match networks, presented at CVPR 2018 (also IJCV 2020). The key is enforcing alignment ...
In contrast to humans, neural networks tend to quickly forget previous tasks when trained on a new one (without revisiting data from previous tasks). In our recent ICPR 2018 paper we propose the rotated elastic weight consolidation (REWC) method to ...

Short bio

I am a researcher (Ramón y Cajal fellow) with the Video Processing and Understanding Lab of the Universidad Autónoma de Madrid. From 2017 to 2023 I was a senior researcher with the Computer Vision Center. From 2012 to 2016 I was with the Institute of Computing Technology (ICT) of the Chinese Academy of Sciences (CAS) in Beijing (China). Previously, I worked with Mitsubishi Electric R&D Centre Europe in Guildford, United Kingdom, and with the Universidad Autónoma de Madrid (UAM), where I received my Ph.D.