↓

BeWater: Effective Protesters Navigate Watersheds in Street Networks

Guillaume Moinard, Matthieu Latapy

In the 25th International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS2026)

During social movements, protesters need to gather with limited communication means and limited knowledge other than what they observe in their direct surroundings. We propose BeWater, a fully distributed walking protocol that achieves gathering thanks to city information like street length, number of restaurants, number of lanes, or street names. Even though using only one of these observables performs poorly, we show that combining them in more advanced tactics rapidly leads to groups of significant sizes. To do so, our work leverages OpenStreetMap data to perform experiments on several real-world cities.

arxiv url to come.

Multi-relational Community Detection in Social Platforms Using Graph Neural Networks

Nouamane Arhachoui, Vincent Gauthier, Anastasios Giovanidis, Lionel Tabourier

In France International Conference on Complex Systems (FrCCS), 2025

We propose a method to detect communities in multi-relational networks, based on a graph neural network pipeline. The method allows to target areas where communities are consensual over the different modes of the network, which are processed as different networks in the pipeline. This is done by combining the outcomes of multiple simple Graph Neural Networks, applied on each of the graphs representing different forms of interactions between users of the social platform. The method is validated on a synthetic benchmark, as a first step for further improvements. In particular, the flexible architecture of the pipeline allows to swap its subparts and create variants of community detection.

Download

Trivial Graph Features and Classical Learning are Enough to Detect Random Anomalies

Matthieu Latapy and Stephany Rajeh

In IEEE International Conference on Data Mining (ICDM), 2025

Detecting anomalies in link streams that represent various kinds of interactions is an important research topic with crucial applications. Because of the lack of ground truth data, proposed methods are mostly evaluated through their ability to detect randomly injected links. In contrast with most proposed methods, that rely on complex approaches raising computational and/or interpretability issues, we show here that trivial graph features and classical learning techniques are sufficient to detect such anomalies extremely well. This basic approach has very low computational costs and it leads to easily interpretable results. It also has many other desirable properties that we study through an extensive set of experiments. We conclude that detection methods should now target more complex kinds of anomalies.

arxiv url to come

Improving flocking behaviors in street networks with vision

Guillaume Moinard, Matthieu Latapy

In International Conference on Complex Networks. Cham: Springer Nature Switzerland, 2025 (CompleNet2025)

We improve a flocking model on street networks introduced in a previous paper. We expand the field of vision of walkers, making the model more realistic. Under such conditions, we obtain groups of walkers whose gathering times and robustness to break ups are better than previous results. We explain such improvements because the alignment rule with vision guaranties walkers do not split into divergent directions at intersections anymore, and because the attraction rule with vision gathers distant groups. This paves the way to a better understanding of events where walkers have collective decentralized goals, like protests.

https://arxiv.org/pdf/2505.21585

Co-evolution of groups and opinions in an agent-based model

Duncan Cassells, Antoine Vendeville, Lionel Tabourier, Pedro Ramaciotti

PLoS One, e0338486, 20 (12), 2025

This study explores a model for the co-evolution of opinions and groups, related to opinion polarization and group identity in opinion dynamics. While traditional models focus on pairwise interactions, we incorporate the notion of groups thereby allowing agents to identify other agents as in-group or out-group. By modifying key parameters we examine how understanding interactions as in-group or out-group affects the convergence or divergence of opinions. Our findings reveal that attitude towards out-group plays a leading role in the polarization of such systems, while in-group interactions can temper extreme opinion shifts or even fragment groups. This model offers an understanding of the complex interplay between group identification and polarization, providing a contribution to ongoing debates about segregation and sectarianism in public and private spheres.

Download

PIQ – Soutien à la recherche à risque et à impact dans le numérique

Guillaume Chelius (INRIA)

Jeudi 22 janvier 2026 à 14h en salle 26-00/534

Lien vers PIQ

Guillaume Chelius, directeur du Programme Inria Quadrant (PIQ), viendra nous parler de ce programme très spécifique de soutien aux démarches de recherche à risque et à impact dans le numérique. Le programme, ouvert depuis fin 2024, accompagne et finance des scientifiques désireux d’engager des projets de recherche à risque et à impact dans le domaine des sciences et technologies du numérique, de ses fondements à ses usages. PIQ s’adresse à tous les scientifiques issus de l’ensemble les établissements publics de l’Enseignement Supérieur et de la recherche, chercheurs, enseignants chercheurs, ingénieurs de recherche. PIQ est ouvert à toute la communauté scientifique académique. PIQ accompagne des projets orientés « porteurs », de durée variable, de six mois à plus de trois ans, et sur des budgets pouvant aller de quelques dizaines milliers d’euros à plus d’un million d’euros. Aucun format n’est imposé. Durant cette session, Guillaume nous présentera le positionnement particulier du programme, la manière dont il s’empare de ces notions de risque et d’impact, ainsi que ses modalités spécifiques d’opération, son processus de candidature et l’ingénierie projet qu’il déploie. Les actualités du programme seront abordées, notamment à travers des projets accompagnés.

Subvertir en perturbant les flux, de Barbe noire à ‘Bloquons tout !’

Dominique Pinsolle, Historien (Université Bordeaux Montaigne)

Jeudi 18 décembre 2025 à 14h en salle 26-00/534

Depuis une vingtaine d’années, la pratique du blocage prend une place croissante dans les mouvements sociaux à l’échelle internationale, des mobilisations altermondialistes au mouvement pro-démocratie à Hong Kong, en passant par les Gilets Jaunes et l’appel « Bloquons tout ! » en France. Si ce mode d’action tranche par sa nouveauté, il s’inscrit également dans une longue histoire de subversion par la perturbation des flux, dont on trouve déjà des manifestations au XVIIe siècle. Cet exposé étudiera la manière dont ont progressivement émergé des tactiques fondées sur l’idée que le pouvoir devait être attaqué dans ses réseaux techniques et ses infrastructures, du XIXe siècle à nos jours.

Responsible digital developments in research and education: where do we start?

Daphné Tuncer (Laboratoire Ville Mobilité Transport, Ecole Nationale des Ponts et Chaussées, IPP)

Jeudi 10 Juillet 2025 à 14h00 en salle 26-00/428

Slides

Several recent initiatives have been proposing new directions for research practices and their operations in the computer science community, from updated codes of conduct that clarify the use of AI-assisted tools to the inclusion of ethical statements and the organisation of working groups on the environmental footprint of digitalisation. In this talk, we will discuss what frameworks are needed to help describe the sense of engagement and accountability with which the practitioner of a computing-related area may be confronted, and guide one to think about the incidence of technical realisations beyond techno-centric contributions. We will share examples from current research projects, including the recent organisation of a session at the 2025 Internet Governance Forum, and teaching practices to illustrate some concrete scenarios.

Inference of multi-dimensional political positions of online users and web

Antoine Vendeville, Post-doctorant (Médialab, Sciences-Po)

Jeudi, 12 Juin 2025 à 11h en salle 24-25/405

Several recent initiatives have been proposing new directions for research practices and their operations in the computer science community, from updated codes of conduct that clarify the use of AI-assisted tools to the inclusion of ethical statements and the organisation of working groups on the environmental footprint of digitalisation. In this talk, we will discuss what frameworks are needed to help describe the sense of engagement and accountability with which the practitioner of a computing-related area may be confronted, and guide one to think about the incidence of technical realisations beyond techno-centric contributions. We will share examples from current research projects, including the recent organisation of a session at the 2025 Internet Governance Forum, and teaching practices to illustrate some concrete scenarios.

Scalable Algorithms to Measure User Influence in Social Networks

Nouamane Arhachoui, Esteban Bautista, Maximilien Danisch, Anastasios Giovanidis, Lionel Tabourier

Social Network Analysis and Mining Applications in Healthcare and Anomaly Detection, pp. 63-92, 2024

Measuring user influence in social networks is crucial for a variety of applications. While traditional centrality metrics evaluate structural graph importance, a more recent metric known as the ψ-score takes into account users’ posting and re-posting activities to provide richer information. The ψ-score is a powerful tool that generalizes PageRank for non-homogeneous node activity. However, for large datasets with N users, it becomes computationally expensive, requiring solving N linear systems of N equations. To tackle this issue, we propose three new scalable algorithms that can quickly approximate the ψ-score. The Power-ψ and Push-ψ algorithms are based on a novel equation that shows it is sufficient to solve one system of equations of size N to calculate the ψ-score. These algorithms take advantage of the fact that the solution of such a system can be recursively and distributedly approximated. Consequently, the ψ-score, which summarizes the nodes’ structural and behavioral information, can be computed as quickly as PageRank. The third proposed algorithm is Push-NF. Despite aiming to solve all N systems to extract additional information on the information dynamics, it still manages to converge to the accurate user ranking faster than the current state-of-the-art alternative. To validate the effectiveness of our proposed algorithms, we release them as an open-source Python library and test them on various real-world datasets.

Download

On the role of diffusion dynamics on community-aware centrality measures

Stephany Rajeh , Hocine Cherifi

Rajeh S, Cherifi H (2024) On the role of diffusion dynamics on community-aware centrality measures. PLoS ONE 19(7): e0306561.

https://doi.org/10.1371/journal.pone.0306561

Theoretical and empirical studies on diffusion models have revealed their versatile applicability across different fields, spanning from sociology and finance to biology and ecology. The presence of a community structure within real-world networks has a substantial impact on how diffusion processes unfold. Key nodes located both within and between these communities play a crucial role in initiating diffusion, and community-aware centrality measures effectively identify these nodes. While numerous diffusion models have been proposed in literature, very few studies investigate the relationship between the diffusive ability of key nodes selected by community-aware centrality measures, the distinct dynamical conditions of various models, and the diverse network topologies. By conducting a comparative evaluation across four diffusion models, utilizing both synthetic and real-world networks, along with employing two different community detection techniques, our study aims to gain deeper insights into the effectiveness and applicability of the community-aware centrality measures. Results suggest that the diffusive power of the selected nodes is affected by three main factors: the strength of the network’s community structure, the internal dynamics of each diffusion model, and the budget availability. Specifically, within the category of simple contagion models, such as SI, SIR, and IC, we observe similar diffusion patterns when the network’s community structure strength and budget remain constant. In contrast, the LT model, which falls under the category of complex contagion dynamics, exhibits divergent behavior under the same conditions.

Download

Fast and Robust Flocking of Protesters on Street Networks

Guillaume Moinard, Matthieu Latapy

In: Luca Maria Aiello, Tanmoy Chakraborty, Sabrina Gaito (eds) 16th International Conference, ASONAM 2024, Rende, Italy, September 2–5, 2024, Proceedings, Part IV

https://doi.org/10.1007/978-3-031-78538-2

We propose a simple model of protesters scattered throughout a city who want to gather into large and mobile groups. This model relies on random walkers on a street network that follow tactics built from a set of basic rules. Our goal is to identify the most important rules for fast and robust flocking of walkers. We explore a wide set of tactics and show the central importance of a specific rule based on alignment. Other rules alone perform poorly, but our experiments show that combining alignment with them enhances flocking, and that obtained groups are then remarkably robust.

Download

La vérité sur la blockchain

Pablo Rauzy, Professeur associé à up8

Mardi, 21 Janvier 2025 à 11h en salle 24-25-405

On entend de plus en plus parler de nouvelles technologies telles que les « cryptomonnaies », le « métavers », les « NFT », ou encore le « web3 », et celles-ci sont invariablement présentées comme des innovations incontournables du monde de demain, sans que ne soit jamais vraiment expliqué ni pourquoi ni comment… sauf une chose : c’est grâce à « la blockchain » ! En plus de ces nouvelles technologies, « la blockchain » est censée également révolutionner certaines pratiques existantes : par exemple la certification de documents (notariat, diplômes) ou la traçabilité (supply chain, agro-industrie), et parfois même, la démocratie (vote électronique)…
Mais, en vrai, ça sert à quoi, une blockchain ?
Après avoir rapidement expliqué les bases du fonctionnement d’une blockchain, nous partirons de cet état de fait technique pour se poser plusieurs questions (et y répondre !) : concrètement, ça fait quoi, une blockchain ? dans quelles hypothèses ? et du coup, quelles sont les limites de cette technologie ? mais alors, est-ce que ça résout un problème qui existe dans la vraie vie ?
En conclusion, nous reviendrons sur le caractère d’« innovation de rupture » systématiquement associé à cette technologie, et nous nous questionnerons sur son rôle en pratique, non plus techniquement, mais socialement et politiquement.

Planetary Limits, Anti-Limits in Computer Systems And The Missing Scenarios

Florence Maraninchi, Professeure à l’INP Grenoble

Jeudi, 9 Janvier 2025 à 11h en salle 25-26-105

Research in computer science and computer engineering includes several branches dedicated to the environmental impacts of ICT. Green-ICT consists in improving the performances of ICT itself (software, hardware, communication infrastruture) in order to reduce its impacts; Green-by-ICT promises to reduce the impacts of other sectors thanks to ICT. In this talk we will argue that this is not sufficient. Green-ICT optimizations are often (if not always) synonymous of massive rebound effects. Green-by-ICT is nothing more than a promise, at least until now. Moreover there are intrinsic anti-limits in the design principles that make it difficult, if not impossible, to stay within planetary limits. We should start studying other, less techno-optimistic, scenarios. A somewhat extreme hypothesis is that manufacturing new hardware will stop at some point in the future. We should therefore study the “fading-ICT” scenario, using the abundant ICT resources of today to prepare a future of scarcity.

Décarboner les mobilités urbaines : premiers résultats pour comprendre et favoriser le cyclisme en ville

Hervé Rivano, professeur à l’INSA de Lyon et chef de l’équipe Agora

mercredi 06 Novembre 2024 à 14h en salle 24-25/509

La transition des mobilités urbaines vers des modes décarbonés est un levier majeur face aux enjeux du dérèglement climatique. En particulier, développer le cyclisme urbain fait partie des stratégies déployées par les métropoles. Pour autant, la compréhension du comportement des cyclistes est encore parcellaire et les modes de partage de l’espace public cantonné à une répartition spatiale des voiries. Dans cet exposé, nous présenterons des contributions, issues de la thèse de Lucas Magnana, à l’analyse des comportements et une piste de partage dynamique de la voirie qui s’appuient sur l’analyse de données de mobilité et des techniques d’apprentissage machine. Des perspectives de recherche, dont une part se fera dans le cadre du PEPR Mobidec, conclueront l’exposé.

Network Analysis Applied to Financial Stocks

Ixandra Achitouv

Mardi 05 Novembre 2024 à 11h en salle 24-25/405

Slides

Financial markets exhibit properties of complex systems. By applying network analysis to stock return correlations, I will present the dynamical properties of the network and their correlation with overall market returns. This approach identifies key variables that provide insight into the complex dynamics of stock interactions and the underlying market structure.

Community detection in directed graphs using stationary distribution and hitting times methods

PHAN Thi Ha Duong

jeudi 24 Octobre 2024 à 14h en salle 26-00/534

Community detection has been extensively developed using various algorithms. One of the most powerful algorithms for undirected graphs is Walktrap, which determines the distance between vertices by employing random walk and evaluates clusters using modularity based on vertex degrees. Although several directions have been explored to extend this method to directed graphs, none of them have been effective. In this paper, we investigate the Walktrap algorithm (Pons and Latapy in J Graph Algorithms Appl 10:191–218, 2006) and the spectral method (Newman in Phys Rev E 88:042822, 2013) and extend them to directed graphs. We propose a novel approach in which the distance between vertices is defined using hitting time, and modularity is computed based on the stationary distribution of a random walk. These definitions are highly effective, as algorithms for hitting time and stationary distribution have been developed, allowing for good computational complexity. Our proposed method is particularly useful for directed graphs, with the well-known results for undirected graphs being special cases. Additionally, we utilize the spectral method for these problems, and we have implemented our algorithms to demonstrate their plausibility and effectiveness.

Géopolitique et réseaux maritimes : l’impact de la guerre Ukraine-Russie sur les connections maritimes de l’Ukraine

Marc-Antoine Faure et Barbara Polo

jeudi 04 juillet de 10h à 12h en salle 24-25/509

Les conflits, qu’ils soient politiques, commerciaux ou militaires, affectent les réseaux de transport. Les opérateurs cherchent à éviter les zones les plus tendues en reconsidérant certaines routes. Des liens peuvent être mis à mal dans le cas des tensions géopolitiques locales, qui peuvent avoir un impact global significatif. Cette présentation propose une analyse du réseau maritime de l’Ukraine et identifie les changements dans sa structure en raison du conflit ayant débuté en 2014, avec l’annexion de la Crimée. Le principal objectif est de mesurer et visualiser les principaux changements survenus dans ce réseau depuis 2010 jusqu’à fin 2023, grâce aux données les plus récentes. L’analyse inclut la modélisation du réseau, la représentation du commerce bilatéral et des routes maritimes. Les principaux résultats confirment l’impact majeur du conflit militaire sur la connectivité portuaire, contribuant ainsi à la littérature sur la vulnérabilité des réseaux maritimes.

Chocs et réseaux maritimes : étude comparée de New York, Kobé et New Orleans

César Ducruet

jeudi 04 juillet de 10h à 12h en salle 24-25/509

Cette présentation s’ouvre sur une brève revue de la littérature sur les chocs dans les réseaux (spatiaux), et plus particulièrement dans le cas des réseaux maritimes. L’absence d’études comparatives a motivé l’analyse conjointe de l’impact de l’attaque des Twin Towers à New York (2001), du tremblement de terre à Kobé (1995), et de l’ouragan Katrina à la Nouvelle-Orléans (2004). L’hypothèse majeure est que des mécanismes identiques sont repérables d’un cas à un autre malgré les différences de nature entre ces chocs. Dans les trois cas et comme attendu, une baisse de trafic significative est observée durant le choc, avec des différences en fonction de la spécialisation commerciale des ports (conteneurs, céréales). Au niveau géographique, on constate une hausse de trafic le long de chaque façade maritime à mesure que la distance à l’épicentre augmente, par effet de diversion. Kobé se distingue par une crise plus longue, son trafic de transit ayant été récupéré par le port proche et concurrent de Busan en Corée du Sud, alors en plein essor. En termes de connectivité, les trois ego-networks se caractérisent par une hausse de leur densité suite au choc, soit une perte d’optimalité dans les circulations maritimes régionales.

↑