A. Reiffers
Home
Research
Summary
Publications
Invited Talks
Teaching
Summary

Recent Research

Safe and robust algorithms in distributed systems

In this research line, we are developping and analyzing new distributed algorithms, based on stochastic approximations and reinforcement learning, to solve resource allocations problems, drift detection and eigenvectors computation.

Decentalized Estimation of Eigenvalues.

  • Estimating Fiedler value on large networks based on random walk observations. Alexandre Reiffers-Masson, Thierry Chonavel, Yezekael Hayel.

Drift detection in time-series.

  • Worst-case drift detection of sensor networks: performances and algorithms. Alexandre Reiffers-Masson.

Distributed Resource Allocations

  • A Backpropagation Approach for Distributed Resource Allocation. Alexandre Reiffers-Masson, Nahum Shimkin, Daniel Sadoc Menasche, Eitan Altman.

  • Utility maximisation in the Coordinator-less IOTA Tangle. Mathilde Jay, Ambre Mollard, Ye Sun, Ruyi Zheng, Isabel Amigo, Alexandre Reiffers-Masson, Santiago Ruano Rincón.

Nudges for Inducing Prosocial Behavior

Information Design

In this work, we investigate, by designing the reputation benefit, how much it is possible to increase the prosocial behavior of a set of agents. Our framework is based on a mean-field Bayesian game.

  • Reputation-Based Information Design for Inducing Prosocial Behavior. Alexandre Reiffers-Masson, Rajesh Sundaresan

Opinion shaping

The goal is to get agents to adopt more pro-social behaviour by targeting them through social advertising. However the topology of the social network is unknown, but we have partial observations while monitoring some agents. We have proposed “reinforcement learning” algorithms for the dissemination/adoption of practices or opinions in a network where topology is unknown.

  • Opinion shaping in social networks using reinforcement learning. Vivek Borkar, Alexandre Reiffers-Masson

Study of social networks through stochastic processes and game theory

In this line of research, we applied the game theory and stochastic processes to model the competition over popularity and visibility in social networks.

Algorithmic contributions

The topic here is to use telecommunication modeling tools to model and shape the timeline in social networks. Typical results are the design of Qoe-enhancing algorithms and the evidence of convergence of these algorithms towards the optimal solution.

  • Posting behavior in Social Networks and Content Active Filtering.Alexandre Reiffers-Masson, Eitan Altman and Yezekael Hayel. International Workshop on Dynamics in Networks (DyNo2015), in conjunction with IEEE/ACM ASONAM 2015, 2015.

  • Timelines are Publisher-Driven Caches: Analyzing and Shaping Timeline Networks. Alexandre Reiffers-Masson, Eduardo Hargreaves, Eitan Altman, Wouter Caarls, Daniel S. Menasche. ACM Performance Evaluation Review (NetEcon Special issue) 2017.

  • Posting behaviour Dynamics and Active Filtering for Content Diversity in Social Networks. Alexandre Reiffers-Masson, Yezekael Hayel, Eitan Altman. IEEE Transactions on Signal and Information Processing over Networks 2017.

  • A Generalized Fractional Program for Maximizing Content Popularity in Online Social Networks. Alexandre Reiffers-Masson, Yezekael Hayel, Eitan Altman and Guillaume Marrel. FAB, in conjunction with IEEE/ACM ASONAM 2018, 2018.

Game theory contributions

The topic here is the study of competition over popularity using game theoretical tools. We are able to study the properties of Nash’s equilibria, of the price of anarchy and to obtain a better understanding of the possible problems induced by a competition for popularity.

  • Game theory approach for modeling competition over visibility on social networks. Alexandre Reiffers-Masson, Yezekael Hayel, Eitan Altman. Communication Systems and Networks (COMSNETS), 2014.

  • Differential Games of Competition in Online Content Diffusion. Francesco De Pellegrini, Alexandre Reiffers-Masson, Eitan Altman. IFIP Networking, 2014.

  • A Time and Space Routing Game Model applied to Visibility Competition on Online Social NetworksAlexandre Reiffers-Masson, Eitan Altman, Yezekael Hayel. NetGCOOP 2014: International conference on NETwork Games, COntrol and OPtimization, 2014.

Data extraction/analysis contributions

We have studied how to estimate the popularity and visibility in social networks. We also studied the bias in the filtering algorithm.

  • A Study of YouTube recommendation graph based on measurements and stochastic tools.Yonathan Portilla, Alexandre Reiffers-Masson , Eitan Altman and Rachid El-Azouzi. 3rd International workshop on Big Data and Social Networking Man- agement and Security, in conjunction with IEEE/ACM UCC 2015, 2015.

  • Fairness in online social network timelines: Measurements, models and mechanism design. Hargreaves, E., Agosti, C., Menasché, D., Neglia, G., Reiffers-Masson, A., & Altman, Performance Evaluation 129, 15-39

  • Fairness in Online Social Network Timelines: Measurements, Models and Mechanism Design. Hargreaves, E., Agosti, C., Menasché, D., Neglia, G., Reiffers-Masson, A., & Altman, E.IFIP Performance 2018, 2018.

  • Hargreaves, E., Agosti, C., Menasché, D., Neglia, G., Reiffers-Masson, A., & Altman, E. Biases in the Facebook News Feed: a Case Study on the Italian Elections. FOSINT-SI, in conjunction with IEEE/ACM ASONAM 2018, 2018.