Information about the coronavirus outbreak, and how it affects me, and my teaching activities : please read this webpage : coronavirus.en.html, I am updating it regularly.
I am Lilian Besson, professor of Computer Science at ENS Rennes, and former student in Mathematics and Computer Science at ENS de Cachan. I am a passionate programmer, an open-source enthusiast, and I did research for my PhD thesis, in machine learning, learning theory and cognitive radio. I also love to cook and meet people, to travel and exchange, to bike or hike, and I am chaotic-good even IRL. Welcome to my website.
PhD thesis (2016-19)¶
For my Ph.D., my research is in applied machine learning, focused on low-cost online learning algorithm with limited feedback (bandit feedback), mainly applied to cognitive radio problems for Opportunistic Spectrum Access and setting up reliable network access protocol for the future Internet of Things networks. By studying and applying classical and recent Multi-Armed Bandit algorithms to carefully designed radio models, we are able to prove some performance guarantees, both numerically in simulations and theoretically with statistical proofs.
Analyse non asymptotique d’un test séquentiel de détection de ruptures et application aux bandits non stationnaires (in French), by L. Besson and E. Kaufmann. GRETSI 2019 conference, Lille, France, August 2019, HAL-02152243.
Decentralized Spectrum Learning for IoT Wireless Networks Collision Mitigation, by C. Moy and L. Besson. 1st ISIoT workshop at the DCOSS 2019 conference, Santorini, Greece, May 2019, HAL-02144465.
Upper-Confidence Bound for Channel Selection in LPWA Networks with Retransmissions, by R. Bonnefoi, L. Besson, J. Manco-Vasquez and C. Moy. 1st MOTIoN workshop at WCNC (Wireless Communication and Networks Conference), Marrakech, Morrocco, January 2018, HAL-02049824.
GNU Radio Implementation of MALIN: « Multi-Armed bandits Learning for Internet-of-things Network », by L. Besson, R. Bonnefoi, C. Moy. IEEE WCNC (Wireless Communication and Networks Conference), Marrakech, Morrocco, April 2019, HAL-02006825.
MALIN: « Multi-Arm bandit Learning for Iot Networks » with GRC: A TestBed Implementation and Demonstration that Learning Helps, by L. Besson, R. Bonnefoi and C. Moy. Demonstration presented at ICT (International Conference on Communication), Saint-Malo, France, June 2018. Cf. YouTu.be/HospLNQhcMk and poster.
Aggregation of Multi-Armed Bandits learning algorithms for Opportunistic Spectrum Access, by L. Besson, E. Kaufmann and C. Moy. IEEE WCNC, Barcelone, Spain, April 2018, HAL-01705292.
Multi-Armed Bandit Learning in IoT Networks and non-stationary settings, by R. Bonnefoi, L. Besson, C. Moy, E. Kaufmann and J. Palicot. CrownCom (Conference on Cognitive Radio Oriented Wireless Networks), Lisboa, Portugal, September 2017, HAL-01575419 and poster, best paper award.
Decentralized Spectrum Learning for Radio Collision Mitigation in Ultra-Dense IoT Networks: LoRaWAN Case Study and Measurements, by C. Moy, L. Besson, G. Delbarre and L. Toutain, July 2019. Submitted for a special volume of the Annals of Telecommunications journal, on « Machine Learning for Intelligent Wireless Communications and Networking ».
The Generalized Likelihood Ratio Test meets klUCB: an Improved Algorithm for Piece-Wise Non-Stationary Bandits, by L. Besson and E. Kaufmann, February 2019. See this page for the code and more details, HAL-02006471.
SMPyBandits: an Open-Source Research Framework for Single and Multi-Players Multi-Arms Bandits (MAB) Algorithms in Python, by L. Besson, in development since Octobre 2016 and still active, HAL-01840022. Code on GitHub.com/SMPyBandits (about 40000 lines), documentation on SMPyBandits.rtfd.io.
Works that need more writing, still in progress for a new submission:
What Doubling-Trick Can and Can’t Do for Multi-Armed Bandits, by L. Besson and E. Kaufmann, September 2018, HAL-01736357.
A Note on the Ei Function and a Useful Sum-Inequality, by L. Besson, February 2018, HAL-01847480.
2nd research M.Sc. (2015-16)¶
For the 1st trimester (Fall 2015):
Parcimonie and Compressed Sensing : « Random factorization for low-rank matrices » (« finding structure with randomness »),
Probabilistic Graphical Models : « Hidden semi-Markov Models » (comparison to Hidden Markov Models and Gaussian Mixture Models),
For the 2nd trimester (Spring 2016):
2nd M.Sc. research internship (2016)¶
theoritical functional analysis, applied to inverse optimization problem, mainly appearing in medical imaging.
april 2016 to August 2016 (research internship in applied mathematics);
1st MSc research internship (2013)¶
« Towards modularity for planning and robot programs verification »;
BSc internship (2012)¶
« Finite volumes method on nVidia graphic cards, applied to solve the compressible Euler problem »;
Math internship at CMLA (Centre des mathématiques et de leurs applications, ÉNS de Cachan math lab research), 5 months (February 2012 to July 2012).
General study of numerical solvers for differential equations and partial differential equations. Liner solver, first and second order, 1 2 and 3 D, with the VFFC method. Numerical simulation, sequential using :C: and VTK, and parallel using nVidia CUDA. Interactive 2D simulation with openGL.
On my web page, the bachelor thesis, in French.