# Alex Nowak-Vila

Since September 2018, I am a PhD student in Machine Learning at INRIA and Ecole Normale Supérieure in the SIERRA project-team in Paris, France. My advisors are Francis Bach and Alessandro Rudi .
My research focuses on structured output prediction, which is the branch of supervised learning that deals with structured outputs such as sequences, graphs, permutations, etc. In particular, I am interested in the statistical and computational properties of predicting structured data such as consistency, sample efficiency and complexity of training and inference.
I hold a master degree in Machine Learning from Ecole Normale Supérieure Paris-Saclay and a double-bachelor in Mathematics and Engineering Physics from Universitat Politècnica de Catalunya in Barcelona.
More information in my resume.

**Email:** alex [dot] nowak-vila [at] inria [dot] fr. **Physical address:** INRIA Paris, 4th floor, Office C416, 2 rue Simone Iff, 75012 Paris.

## Research Interests

- Structured output prediction.
- Statistical learning theory.
- Kernel methods.
- Convex optimization.

## Publications

- Eboli, T*., Nowak-Vila, A*., Sun, J., Bach, F., Ponce, J, Rudi, A. -
** Structured and Localized Image Restoration. **preprint. [pdf]
- Nowak-Vila, A., Bach, F., and Rudi, A. -
** Consistent Structured Prediction with Max-Min Markov Networks. ***Proceedings of the International Conference on Machine Learning (ICML), 2020.* [pdf] [slides] [code]
- Nowak-Vila, A., Bach, F., and Rudi, A. -
** A General Theory for Structured Prediction with Smooth Convex Surrogates.** preprint. [pdf]
- Nowak-Vila, A., Bach, F., and Rudi, A. -
** Sharp Analysis of Learning with Discrete Losses.** *Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), 2019.* [pdf] [poster]
- Nowak-Vila, A., Folque, D., and Bruna, J. -
** Divide and Conquer Networks.** *International Conference on Learning Representations (ICLR), 2018.* [pdf] [code] [poster]
- Nowak-Vila, A., Villar, S., Bandeira, A., Bruna, J. -
** Revised Note on Learning Algorithms for Quadratic Assignment with Graph Neural Networks.** * PADL Workshop at ICML 2017. * [pdf] [code] [poster]

## Software

Link to my Github.