Nicolas Carrara - Ph.D. candidate

Nicolas Carrara

[surname].[name]@[inria.fr|protonmail.com]

Inria Lille – Nord Europe, équipe SequeL Parc Scientifique de la Haute-Borne 40 avenue Halley 59650 Villeneuve d’Ascq, FRANCE

I'm a postdoctoral fellow in machine learning at the University of Toronto. I'm particularly interested in deep reinforcement learning. Robotics, dialogue systems, autonomous driving and video games are my favorite applications. I was a member of the team SequeL at INRIA and NADIA team at Orange Labs.

This website has been generated with node.js, express.js and server-side templating handlebars.js using this data file as content. CSS, HTML and client-side JS from this template.

My bibtex.

See my one page resume

Some of my internet fingerprints (please note the Github repository is basically from my post Msc era while Bitbucket is from my pre Msc era):

[last update: May 2020]

Research Experience

Reviewing




Postdoctoral Fellow

University of Toronto

Deep Reinforcement Learning for Traffic Control.


Team: D3M lab
Supervisors: Assistant Pr. Scott Sanner Pr. Baher Abdulhai
March 2020 - present

PHD STUDENT

ORANGE, CRISTAL AND LILLE1

Reinforcement learning for dialog systems optimization with user adaptation. http://ncarrara.fr/others/thesis-nicolas-carrara.pdf


Team: Sequel (INRIA) NADIA (Orange Labs)
Supervisors: Pr. Olivier Pietquin (thesis director) Dr. Romain Laroche Dr. Tanguy Urvoy Dr. Jean-Léon Bouraoui
October 2015 - December 2019

Research intern

Inria

Master thesis : automated planning under uncertainty with multiple objectives.


Team: LARSEN
Supervisors: Dr. Olivier Buffet Dr. Vincent Thomas
March 2015 - August 2015

Research Intern

Inria

Improvements of a new dynamic neural field model: Randomly Spiking Dynamic Neural Fields (RSDNF).


Team: CORTEX
Supervisors: Pr. Bernard Girau Dr. Benoît Chappet de Vangel
June 2014 - August 2014

Research project

Inria

Predicting user behavior on the web.


Team: KIWI
Supervisors: Dr. Samuel Nowakowski
January 2014 - May 2014

Research Intern

Inria

Adding a model for dynamic neural field. Basic image processing.


Team: CORTEX
Supervisors: Pr. Bernard Girau Dr. Benoît Chappet de Vangel
June 2013 - August 2013

Teaching Experience

Teaching Assistant - RLSS

Reinforcement Learning Summer School

Helping students during lab sessions of RLSS 2019.


Promotion: N/A

June 2019 - June 2019

Temporary Research and Teaching Attaché - Computer science

Université Lille 3

Teaching python, regular expressions, nodejs etc.


Promotion: Licence MIASHS

September 2018 - September 2019

TEACHER - REINFORCEMENT LEARNING

UNVIVERSITÉ LILLE 1

Lectures and lab sessions of reinforcement learning for computer science MOCAD master.


Promotion: Master MOCAD

January 2018 - February 2018

TEACHING ASSISTANT - WEB DEVELOPMENT

UNIVERSITÉ LILLE 1

Lab sessions of html/css/php/javascript to SIAD master students.


Promotion: Master SIAD

January 2017 - May 2017

TUTOR

UNIVERSITÉ DE LORRAINE

Tutor for the computer science part of the 1st year of university.


Promotion: Licence 1 d'informatique

September 2014 - January 2015

Education

LXMLS 2017

SUMMER SCHOOL
Lisbon Machine Learning School for Natural Language processing applications.

Rank: N/A
July 2017-July 2017

UNIVERSITÉ DE LORRAINE

RESEARCH MASTER'S DEGREE
IPAC : Machine learning, Data mining, Robotics, Image recognition.

Rank: Head of the class
2014-2015

Publications

Budgeted Reinforcement Learning in Continuous State Space [url]
Carrara, Nicolas and Leurent, Edouard and Laroche, Romain and Urvoy, Tanguy and Maillard, Odalric and Pietquin, Olivier
Neural Information Processing Systems (NeurIPS2019)
2019

Reinforcement learning for Dialogue Systems optimization with user adaptation. [url]
Carrara, Nicolas
Ph.D. thesis
2019

Safe transfer learning for dialogue applications [url]
Carrara, Nicolas and Laroche, Romain and Bouraoui, Jean-Léon and Urvoy, Tanguy and Pietquin, Olivier
International Conference on Statistical Language and Speech Processing (SLSP 2018)
2018

A Fitted-Q Algorithm for Budgeted MDPs [url]
Carrara, Nicolas and Laroche, Romain and Bouraoui, Jean-Léon and Urvoy, Tanguy and Pietquin, Olivier
Workshop on Safety, Risk and Uncertainty in Reinforcement Learning, Uncertainty in Artificial Intelligence (UAI 2018)
2018

A Fitted-Q Algorithm for Budgeted MDPs [url]
Carrara, Nicolas and Laroche, Romain and Bouraoui, Jean-Léon and Urvoy, Tanguy and Pietquin, Olivier
European Workshop on Reinforcement Learning (EWRL 2018)
2018

Online learning and transfer for user adaptation in dialogue systems [url]
Carrara, Nicolas and Laroche, Romain and Pietquin, Olivier
Joint special session on negotiation dialog, Semantics and Pragmatics of Dialogue (SemDial 2017)
2017

Projects

  • An api to use Pydial as a Gym Environment [code].
  • An api to make custom scenarios programmatically on Age of Empire Definitive Edition [code].
  • An Arduino project to create an autonomous UAV from scratch with Reinforcement Learning [code].
  • Working around with youtube-dl and others stuff ... [demo].
  • Parse your google scholar page [code].

Ressources

MIASHS

Fonctionnement du système informatique Projet informatique Internet et base de donnée. (Notes)

UE10

Traitement de texte et tableaur Traitement automatique de corpus (Notes)