The current research of my team focuses on the design of interpretable deep learning models for computer vision, data mining and natural language processing. We also work on the understanding of deep models via explanation methods and theoretical bounds, and on the efficient application of deep learning for big data.

In the domain of computer vision: we design deep neural networks using deep unfolding principles and address image/video processing and analysis tasks, including super-resolution, denoising, robust principal component analysis (for anomaly detection and foreground-background detection) and image classification and segmentation. We work with various imaging modalities, including RGB, infrared, multi-/hyper-spectral, and X-ray imaging.

In the domain of data mining: our research focuses on deep-learning models for matrix/tensor factorisation and geometric deep-learning models for graph data processing and analysis. We consider various applications including recommender systems and hyper-local air quality estimation in cities.


In the domain of Natural Language Processing (NLP): Our research focuses on solving core problems, using either classical machine learning or deep learning based models, and imposing structure into raw text. Recent problems we focus on include fake news detection, fact verification, and event detection from social media sources.