About me
You can contact me at:
I am a 23-year old Computer Science student at Ecole Normale Supérieure de Rennes, France. I am mainly interested in Automated Planning, Operations Research and theoretical CS - especially Complexity Theory and Logics.
I am extremely fond of algorithmics and programming, and I enjoy testing myself in various competitions, especially in team.
I enjoy arts, be it literature, graphical arts or music. I have been playing the guitar for a few years now. In my spare time, I also solve a few Rubik's Cubes.
Close
Résumé
Education
Click on a box for more details
Master 2 of Research - Science Informatique Fondamentale -
University of Rennes I
Classes taken:
First Semester:
- Category Theory - Class link
- Data Mining and Visualization
- Game Theory and Application to Networks
- Advanced Probabilistic Data Analysis and Modeling - Class link
- High-Dimensional Learning
- Deep Learning for Vision
- Models and Algorithms for Distributed Computing
- Pedagogy - Unplugged Computing
During this year, I pursued two different internships in foreign universities, in order to gain research experience. See below for more information.
Master 1 of Research - Science Informatique Fondamentale, co-accredited with
University of Rennes 1.
Average: 15.74/20
Classes taken:
First Semester:
- Automata-based Model Checking
- Complexity Theory
- Syntax-Driven Compiling
- Signal Processing
- Solvers Principles and Architecture - Class link
- Formal Analysis and Design of Programs - Introduction to Isabelle/HOL
- Master Project
- English
- Initiation to Research
Second Semester:
- Information Theory
- Static Analysis for Programs Optimization
- Constraint and Logic Programming
- Artificial Intelligence - Introduction to Machine Learning and Data Mining
- Advanced Databases
- Epistemology
- Master Project
- English
- Initiation to Research
Licence Informatique Fondamentale, co-accredited with
University of Rennes 1.
Average of 15.038/20, Mention Bien
Ranked 4th out of 35
Classes taken:
First Semester:
- Formal Languages and Computability
- Introduction to Lebesgue's Integral and Measure Theory
- Convex Optimisation
- Applied Algebra: Error-Correcting Codes
- Algorithmics 1
- Programming 1
- Systems and Networking
- Introduction to Computer Security
- English
Second Semester:
- Algorithmics 2
- First-Order Logic
- Probabilities and Applied Statistics
- Programming 2
- Computer Architectures
- Images Classification
- Cryptography
- Pedagogy
- Initiation to Research
- English
French preparatory program to the Grandes Ecoles.
Intensive preparation in Mathematics, Physics and Computer Science.
With English, Literature and Philosophy classes.
French high school diploma with highest distinction.
Average of 19/20
Main Topics: Mathematics, Physics, Biology.
Research Experience
Under the supervision of Prof. Thomas Bolander, PhD. Technical University of Denmark.
Study on the parameterized complexity of the Dynamic Belief Update problem, a variant of the model checking problem for Dynamic Epistemic Logic. Our joint work lead to a publication. [pdf]
Under the supervision of Prof. Hector Geffner, PhD. Universitat Pompeu Fabra.
During this internship, I tried to bridge the gap between Planning Width, Reinforcement Learning and Correlation Complexity.
Under the supervision of Dr. Silvan Sievers, University of Basel.
Merge & Shrink abstractions are useful for finding suitable heuristics for optimal planning. During my internship, I implemented weak-bisimulation based shrinking strategies in Fast Downward.
Under the supervision of Prof. Damien Pellier, PhD, Laboratoire d'Informatique de Grenoble.
Hierarchical Task Network (HTN) Planning is a variant of automated planning, where actions are abstract actions and can be recursively decomposed, in order to form a plan. During my internship, I wrote bencharks for PDDL4J, an HTN planning framework developped in Team MAGMA.
Publications
- Thomas Bolander, Arnaud Lequen. Parametrized Complexity of Dynamic Belief Updates. In Proceedings of Dynamic Logic: New Trends and Applications (DaLí), 2020. To appear in Lecture Notes in Computer Science, 2020. [pdf] [slides]
Languages
English certifications:
- CAE: 198/210
- TOEIC: 990/990
Extra-curricular activities
Close
Projects
Here is a list of a few projects I did, by myself or in team. Clicking on a picture will display additional information about the selected project.
Close
Nowhere-Zero Flows
Nowhere-zero flows (NZF) are related to the common notion of flows, except that they take their values in any finite abelian group. They are a generalization of the k-flow problem, as
formalized by W. Tutte with the 5-flow conjecture.
I come back to this project every now and then. I first implemented a polynomial algorithm that finds a k-flow for k > 3, through its dual problem, 4-coloration.
Later, I developped an algorithm that tries to solves the general problem when possible.
Although my works shows no novel technique, it helped me get familiar with various operations research algorithms.
A (very) quickly written report is available here.
Close
Lisp Interpreter
A quick Lisp interpreter with garbage-collection and dynamic typing analysis, written in C++.
This project was done with two of my Programming 2 classmates.
Close
Delaunay Triangulation and
Voronoï Tesselation
Delaunay triangulation is a classic algorithm, with many interesting applications. The program, done in OCaml, computes such a triangulation for a given graph.
The duality of Delaunay Triangulation and Voronoï Tesselation allowed us to implement an algorithm commuting the later, as shown by the above picture.
This project was done with two of my Programming 1 classmates Victor and Clément.
Close
Basic Raytracing Engine
With access to a few OpenGL functions only, we implemented a basic raytracing engine in C++. Being a first step in computer graphics, it allowed me to get myself familiar with the Phong model.
The program supported basic geometries, such as spheres and planes, as well as a simple reflection model, allowing glossy effects and reflections.
This project was done with two of my Language Theory 1 classmates.
Close
Space Shooter
Simple Java game made with LibGDX. Twin-sticks shooter in space, inspired by Geometry Wars
The game handled a few basic enemies in a successions of rooms. Although it could only boast a few weapons and enemies, the game was made so that it would be easy to add more content.
I planned, at first, to make it easily moddable.
The game was meant to feature some RPG elements and deepter mechanics, but my lack of organization - the project was made in 2013 - and my shallow knowledge in game design and development patterns hampered my progress.
Close
Merge & Shrink Heuristics
An implementation of several weak-bisimulation based shrinking heuristics for the Merge and Shrink framework, implemented in Fast Downward.
It was done during my internship under the supervision of Dr. Silvan Sievers, in the AI Group of the University of Basel, which is in charge of the maintenance of the Fast Downward planning framework.
In addition to implementing the heuristics, I also made an extensive use of Downward Lab, a very convenient tool for the evaluation of planners.
Close
Parameterized Problem Solver
A simple problem in Python meant to facilitate research in parameterized complexity, by finding quickly which sub-problems of a problem with multiple parameters can have their complexity decided with a database of
known reductions, and of known tractable and intractable problems. In addition, problems whose complexity is not yet known can be found, and the impact of finding their complexity is given by the program.
This was done during my internship with Prof. Thomas Bolander, PhD, on a parameterized version of a variant of the model-checking problem for Dynamic Epistemic Logic. As such, the code comes with various examples in that domain.
The code can be found here.
Close
Algorithmic Club Website
Link to the website: Club Algo (in English with French parts)
In order to allow group training for my classmates and me, I set up a website meant for competitive programming. It is an enhanced version of DMOJ, on which it is based.
Built with Django, a Python framework, I extended the website so that it supports:
- Optimisation problems. Problems which are intractable, and whose optimal solution can not be computed in reasonable time. I added this feature in order for us to train for Google HashCode.
- AI competitions. The final of the Prologin contest is a tournament between programs developed to play a game designed for the occasion.
I implemented a functionnality that allows users to upload and compile their program on the website, and make it compete against other players. A visualiser also allows them to understand better
the behavior of their AI and the opponent's one.
I also created a few problems, as well as test cases for a lot of differents problems the site showcases. I also added some tutorials and classes, to help newcomers leap into competitive programming.
Close
Contact
Feel free to use any of the following means to reach me.
Those are my profesionnal details.
|
|
|
Arnaud Lequen
École Normale Supérieure de Rennes
Campus de Ker Lann
35170 Bruz, France |
Close
