Research

I am Bernardino Romera-Paredes, currently Postdoc at University of Oxford, having recently finished my PhD at University College London (UCL). My research is focused on building user-dependent models for affect and expression recognition by combining and extending multilinear and multi-task learning methods. These models have been successfully applied to a variety of scenarios beyond affect recognition, in which data sets can be arranged into meaningful multi-modal structures. This research has led to publications in top-tier machine learning conferences such as NIPS, ICML and AISTATS, as well as interaction based conferences, such as Affective Computing and Intelligent Interaction (ACII) and Face and Gesture (FG). I received the Best Paper Award at ACII 2013 and the Best Paper Runner-up Prize at ICML 2013, among others.

In 2013, I did an internship at Microsoft Research in Redmond, where I investigated about the 3D reconstruction of facial expressions from video captured head-mounted cameras.

Prior to my PhD, I finished with honors Bachelor's Degree in Computer Engineering at the University of Murcia and completed with distinction the MSc Computational Statistics and Machine Learning at UCL.



Publications

2014

  • B. Romera-Paredes, C. Zhang, Z. Zhang: Facial Expression Tracking from Head-Mounted, Partially Observing Cameras. IEEE International Conference on Multimedia & Expo, IEEE ICME 2014. Chengdu, China. [paper] [project page]
  • A. Maurer, M. Pontil, B. Romera-Paredes: An Inequality with Applications to Structured Sparsity and Multitask Dictionary Learning. Conference on Learning Theory, COLT 2014, Barcelona, Spain [paper]



2013

  • B. Romera-Paredes, M. Pontil: A New Convex Relaxation for Tensor Completion. Neural Information Processing Systems, NIPS 2013. Lake Tahoe, USA. [paper] [appendix] [poster]
  • H. J. Griffin, M. S. H. Aung, B. Romera-Paredes, C. McLoughlin, G. McKeowny, W. Currany, N. Bianchi-Berthouze: Laughter Type Recognition from Whole Body Motion. Affective Computing and Intelligent Interaction, ACII 2013. Geneva, Switzerland. Best paper award. [paper]
  • A. Maurer, M. Pontil, B. Romera-Paredes: Sparse coding for multitask and transfer learning. International Conference on Machine Learning, ICML 2013. Atlanta, USA. Best paper runner-up. (Authors have the same contribution) [paper] [appendix]
  • B. Romera-Paredes, H. Aung, N. Bianchi-Berthouze, M. Pontil: Multilinear Multitask Learning. International Conference on Machine Learning, ICML 2013. Atlanta, USA. [paper]
  • M.S.H. Aung, B. Romera-Paredes, A. Singh, S. Lim, N. Kanakam, A. C. de C. Williams, N. Bianchi-Berthouze. Getting rid of pain-related behaviour to improve social and self perception: A Technology-Based Perspective. 14th International Workshop on Image and Audio Analysis for Multimedia Interactive Services, WIAMIS 2013. Paris, France. [paper]
  • B. Romera-Paredes, H. Aung, N. Bianchi-Berthouze: A One-Vs-One Classifier Ensemble with Majority Voting for Activity Recognition. European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2013. Bruges, Belgium. Winner of the ESANN 2013 Competition on Automatic Human Activity Recognition [paper]
  • B. Romera-Paredes, H. Aung, M. Pontil, N. Bianchi-Berthouze, A. C. de C. Williams, P. Watson: Transfer Learning to Account for Idiosyncrasy in Face and Body Expressions. 10th Conference on Automatic Face and Gesture Recognition, IEEE FG 2013. Shangai, China.


2012

  • B. Romera-Paredes, A. Argyriou, A.C.D.C. Williams, N. Berthouze, M. Pontil, Automatic Recognition of Facial Expressions. 14th World Congress on Pain, IASP 2012. Milan, Italy.
  • B. Romera-Paredes, A. Argyriou, N. Bianchi-Berthouze, M. Pontil: Exploiting Unrelated Tasks in Multi-Task Learning. Artificial Intelligence and Statistics, AISTATS 2012. La Palma, Spain. [paper] [poster]


2011

  • B. Romera-Paredes, M. Pontil, N. Bianchi-Berthouze: Leveraging Different Transfer Learning Assumptions: Shared Features, Hierarchical and Semi-Supervised. Challenges in Learning Hierarchical Models. NIPS Workshop 2011. Granada, Spain.
  • H. Meng, B. Romera-Paredes, N. Bianchi-Berthouze: Emotion recognition by two view SVM_2K classifier on dynamic facial expression features. 9th Conference on Automatic Face and Gesture Recognition, IEEE FG 2011, Workshop. Santa Barbara, USA.


2010

  • G. Acampora, G. Fenza, E. Muñoz, B. Romera-Paredes, Mejoras en el uso de Aprendizaje Activo con Árboles Fuzzy: un ejemplo en la toma de decisiones de un Sistema Coordinado [Improvements in Fuzzy Trees by making use of Active Learning: an example of Making Decision in a Coordinate System], XV Spanish Congress on Fuzzy Logic and Technologies, ESTYLF 2010. Huelva, Spain.


2009

  • J. M. Cadenas, C.M. Garrido, E. Muñoz, B. Romera-Paredes, Un Sistema Híbrido de Metaheurísticas en Entornos Dinámicos[An Hybrid Metaheuristic System in Dynamic Environments], VI Spanish Congress on Metaheuristics, Evolutive and Bioinspired Algorithms, MAEB 2009, Málaga, Spain.
Comments