Detection of Weak Signals in Graph-Oriented Data: An Approach Using LLM and DGNN
Role: PhD co-supervisor Dates: 2024 - 2027 Funding: PhD student salary + support Source of funding: Auvalie Innovation Company, ANRT
Collaborators: Ludovic Moncla, Khalid Benabdeslem
Thèse CIFRE
BIT-STABLENET: Extraction of stable relations between entities in crypto-currencies
Role: Project Leader, Post-doc main supervisor Dates: 2023 - 2025 Funding: 2 years of Post-Doc salary Source of funding: Lyon 1 University SENS
Collaborators: -
The objective of this project is to develop methods for the extraction of stable relations between entities in crypto-currencies.
DECOFLO: Detection of dynamic communities in massive link streams
Role: Project Leader, PhD main supervisor Dates: 2023 - 2026 Funding: PhD student salary + support Source of funding: Sahar Company Collaborators: Angela Bonifati
The objective of this project is to develop methods for the detection of dynamic communities in massive link streams.
HYGRAPH: Combining Temporal Graphs and Time Series Into a Single Abstraction
Role: Project member Dates: 2022 - 2025 Website: Link Source of funding: Franco-German call, ANR/DFG Collaborators: PIs: Lyon : Angela Bonifati. Leipzig: Erhard Rahm
The objective of this project is to develop methods for the extraction of stable relations between entities in crypto-currencies.
ExpoXai
Role: PhD main supervisor Dates: 2020 - 2023 Funding: PhD student saraly + support Source of funding: ANRT CIFRE Collaborators: Meersens - ISGlobal (Barcelona)
This project aims at using Explainable AI to better understand the role of exposome on human health. The exposome describes the set of all factors to which
inidivuals are exposed. Understanding which factor has an effect on which person is a major challenge given the quantity and variety of factors. We propose in this project to use state-of-the-art machine learning techniques,
and to use Explainable AI (XAI) to provide explanations on the risks to which individuals are confronted to.
The goal of this project is to study the relation between mosquito and urban environment, in particular in the context of fighting Mosquito-driven viruses. Our contribution consist in the anlysis of correlation networks between different factors impacting mosquito development.
Understanding Network Organization
Role: - Dates: 2010 - ... Funding: None / Many Source of funding: None
I'm interested in the organization of networks: communities, spatial structure, and any other type of organization. I work on this question when I'm not busy on another project.
Past Projects
BITUNAM: BITcoin User Network Analysis and Mining
Role: Project Leader, PhD main supervisor Dates: 2019 - 2023 Funding: 150k€ Source of funding: ANR Collaborators: To be Completed
Automated Urban Vertical Farms represent a new way to produce plants/vegatables, in a fully controlled environment.
The goal of this project is to use data analysis techniques to improve the efficiency of this Industry 4.0 techniques. We will in particular focus on Predictive maintenance and Optimization of plant growing recipes.
INTERASCO : Multipartite Interactions and Consequences for Mosquito Holobiont
The goal of this project is to study the mosquito holobiont, in particular in the context of fighting Mosquito-driven viruses. Our contribution consist in the anlysis of correlation networks inside the holobiont.
Sequential Graph Neural Networks for traffic prediction
The goal of this exploratory project is to initiate a collaboration between reserachers in Machine learning and in the traffic domain in Lyon, in particular on the usage of Graph Neural Networks for traffic prediction.
Automated Discovery of the Organisation of Complex Systems
Role: Project Leader, PhD main supervisor Dates: 2017 - 2021 Funding: 8.5k€ + 1 PhD Student (env. 100k€) Source of funding: Lyon 1 University (BQR)
CDSN (PhD student)
Collaborators: Claire Lesieur (Laboratoire Ampere, ENS de Lyon, Lyon, France)
Complex Systems are systems composed of multiple parts in interaction that cannot be understood by reductionnism. Interactions/Relations between parts of these systems can be represented by complex networks. Several models of networks organisation exist: Communities, Core/Periphery, Low dimensional embedding, etc. In this project, we will explore how one can automatically discover the best model for a given network without a priori structure, using variants of Occam Razor's principle.
Studying properties of networks Embedding
Dates: 2018 Role: Project Leader, Master Internship Co-supervision Funding: 3.5k€ Source of funding: Federation Informatique de Lyon Collaborators: Christine Largeron (Laboratoire Hubert Curien, St Etienne, France)
Graph embedding, or network embedding, is a potential game changer in network analysis. In this exploratory project, we will catch up with state of the arts tools and techniques for network embedding, and propose an embedding preserving the dynamic of networks. We will particularly focus on highly dynamic networks composed of link streams, or instantaneous interactions.
iTRAC : Fraud detection in Bitcoin transaction network
Dates: 2016-2017 Role: Project Member Source of funding: FUI Collaborators: THALES, PAYMIUM
The goal of this project was to use data mining and graph analysis techniques to help discover suspicious activities in the bitcoin transaction network.
VEL'INNOV: Bicycle Sharing Systems, Social appropriations of an Innovative Socio-Technical System
Bicycle Sharing Systems are now ubiquitous in large cities on the planet. My role in this project was to use network analysis and data mining to better understand usages of such systems.