Research

My principal fields of interest:

• Game Theory (Matchings, Voting Systems)

• Large Networks Modeling and Performance Analysis

• Scientometrics (the science of publications)

• Efficient, Unsupervised Natural Language Processing

• Peer-to-peer systems

All my publications are now managed via HAL, the French open archive for research.

Some of my GitHub repositories are related to my research. See them here!

Selected Publications

• Online Stochastic Matching: A Polytope Perspective: Items arrive in a system. Compatible items can be matched and exit the system. How many degrees of freedom can we have without making the system unstable?

• Performance of Balanced Fairness in Resource Pools: A Recursive Approach: In this article, we proposed a simple formula (Theorem 1) that allows us to predict the service rates in a scenario where multiple job types can be treated by multiple servers following some assignment graph.

• Kleinberg’s Grid Unchained: As part of the Kleinberg’s Grid Trilogy, we aim to shed new light on Kleinberg’s model for small world routing using a fast, dedicated simulator. Despite being entirely simulation-centric, this work was published in the Theoretical Computer Science journal.

• Can a Condorcet Rule Have a Low Coalitional Manipulability?: This article on voting systems is centered around a simple result: if low (coalitional) manipulability is sought, then methods that respect the Condorcet criterion should be considered.

• Epidemic live streaming: optimal performance trade-offs: What is the performance (loss and delay) of gossip-like chunk diffusion techniques?

• Self-stabilization in preference-based systems: The PBS framework helps understand the performance of unstructured distributed systems where interactions are driven by participant interests. One key ingredient is acyclicity, a property ensured by node- and edge-based metrics, which provides convergence guarantees.