arXiv subjects

Jean-Philippe Cointet

Publications and source records attributed to Jean-Philippe Cointet.

At least 19 recordsLinked to original sources

arXiv2026

Toward Collective-Centric Evaluation of Preference Inference for Participatory Democracy

To scale up collective decision-making, participatory democracy platforms such as Polis and Remesh enable online deliberation among thousands of participants. However, at this scale, participants cannot review every opinion submitted by others, producing highly sparse voting data that misrepresent patterns of consensus, conflict, and minority support. Platforms therefore increasingly rely on Preference Inference (PI) models to predict missing votes. Yet this automation is not neutral: inferred preferences can artificially amplify, suppress, or reorder existing patterns of support, ultimately reshaping how the outcomes of a deliberation are interpreted. More generally, we lack a systematic understanding of how existing PI methods affect the collective preference landscape. To address this gap, we benchmark several existing PI approaches in this context. Moving beyond conventional user-centric evaluations centered on the accuracy of individual predictions, we introduce a collective-centric evaluation framework that measures whether inferred votes preserve salient properties of the broader preference landscape. We further contribute the largest multilingual dataset of its kind: four consultations spanning over 90k participants, 1M votes, and 22 languages. Our experiments show that models with comparable predictive accuracy can differ substantially in the degree to which they preserve the collective structure. These results demonstrate that accuracy alone is insufficient for evaluating PI in democratic settings. By contributing a novel comprehensive and collective-centric evaluation benchmark for the task of PI, this work aims to support the development of AI systems that scale deliberation without compromising the integrity of its democratic outcomes.

cs.SI↗

arXiv2026

SPOT: An Annotated French Corpus and Benchmark for Detecting Critical Interventions in Online Conversations

We introduce SPOT (Stopping Points in Online Threads), the first annotated corpus translating the sociological concept of stopping point into a reproducible NLP task. Stopping points are ordinary critical interventions that pause or redirect online discussions through a range of forms (irony, subtle doubt or fragmentary arguments) that frameworks like counterspeech or social correction often overlook. We operationalize this concept as a binary classification task and provide reliable annotation guidelines. The corpus contains 43,305 manually annotated French Facebook comments linked to URLs flagged as false information by social media users, enriched with contextual metadata (article, post, parent comment, page or group, and source). We benchmark fine-tuned encoder models (CamemBERT) and instruction-tuned LLMs under various prompting strategies. Results show that fine-tuned encoders outperform prompted LLMs in F1 score by more than 10 percentage points, confirming the importance of supervised learning for emerging non-English social media tasks. Incorporating contextual metadata further improves encoder models F1 scores from 0.75 to 0.78. We release the anonymized dataset, along with the annotation guidelines and code in our code repository, to foster transparency and reproducible research.

cs.CL↗

arXiv2026

Algorithmic Approaches to Opinion Selection for Online Deliberation: A Comparative Study

During deliberation processes, mediators and facilitators typically need to select a small and representative set of opinions later used to produce digestible reports for stakeholders. In online deliberation platforms, algorithmic selection is increasingly used to automate this process. However, such automation is not without consequences. For instance, enforcing consensus-seeking algorithmic strategies can imply ignoring or flattening conflicting preferences, which may lead to erasing minority voices and reducing content diversity. More generally, across the variety of existing selection strategies (e.g., consensus, diversity), it remains unclear how each approach influences desired democratic criteria such as proportional representation. To address this gap, we benchmark several algorithmic approaches in this context. We also build on social choice theory to propose a novel algorithm that incorporates both diversity and a balanced notion of representation in the selection strategy. We find empirically that while no single strategy dominates across all democratic desiderata, our social-choice-inspired selection rule achieves the strongest trade-off between proportional representation and diversity.

cs.CY↗

arXiv2026

Expressing One's Identity Online: Left-Right and cross EU-country variation in self-representation in social media

We examine how social media users from eight European Union (EU) member states express their socio-political identities, focusing on users' online self-presentation and group identity cues conveyed through bios. Our goal is to explore commonalities and differences in topics discussed in social media profiles, across Left-and Right-wing user groups, within and across EU countries. Through a novel approach we map how identity-related discourse varies by country and political orientation, revealing how group identity is expressed within the EU. We find that topics related to democracy, national way of life, and decentralization emerge as particularly divisive, showing considerable variation both within and between EU countries. A subset of topics, which includes education, environmentalism, sustainability, equality, freedom & human rights, and traditional morality, among others, clearly differentiate Left-from Right-leaning user groups. These partisan topics are relevant as they could be leveraged for mobilizing ideological groups and highlight Left-Right identitarian differences at the EU level. Finally, we show that our Left-Right identity similarity metrics reflect aspects of real-world political fragmentation, which are closely aligned to the perceptions of political conflict intensity by country, as measured by the 2022 PEW survey.

cs.SI↗

arXiv2026

Mapping the political landscape from data traces: multidimensional opinions of users, politicians and media outlets on X

Studying political activity on social media often requires defining and measuring political stances of users or content. Relevant examples include the study of opinion polarization, or the study of political diversity in online content diets. While many research designs rely on operationalizations best suited for the US setting, few allow addressing more general political systems, in which users and media outlets might exhibit stances on multiple ideology and issue dimensions, going beyond traditional Liberal-Conservative or Left-Right scales. To advance the study of more general online ecosystems, we present a dataset pertaining to a population of X/Twitter users, parliamentarians, and media outlets embedded in a political space spanned by dimensions measuring attitudes towards immigration, the EU, liberal values, elites and institutions, nationalism and the environment, in addition to left-right and liberal-conservative scales. We include indicators of individual activity and popularity: mean number of posts per day, number of followers, and number of followees. We provide several benchmarks validating the positions of these entities and discuss several applications for this dataset.

cs.SI↗

arXiv2025

Graphically Speaking: Unmasking Abuse in Social Media with Conversation Insights

Detecting abusive language in social media conversations poses significant challenges, as identifying abusiveness often depends on the conversational context, characterized by the content and topology of preceding comments. Traditional Abusive Language Detection (ALD) models often overlook this context, which can lead to unreliable performance metrics. Recent Natural Language Processing (NLP) methods that integrate conversational context often depend on limited and simplified representations, and report inconsistent results. In this paper, we propose a novel approach that utilize graph neural networks (GNNs) to model social media conversations as graphs, where nodes represent comments, and edges capture reply structures. We systematically investigate various graph representations and context windows to identify the optimal configuration for ALD. Our GNN model outperform both context-agnostic baselines and linear context-aware methods, achieving significant improvements in F1 scores. These findings demonstrate the critical role of structured conversational context and establish GNNs as a robust framework for advancing context-aware abusive language detection.

cs.CL↗

arXiv2021

Uncovering the structure of the French media ecosystem

This study provides a large-scale mapping of the French media space using digital methods to estimate political polarization and to study information circuits. We collect data about the production and circulation of online news stories in France over the course of one year, adopting a multi-layer perspective on the media ecosystem. We source our data from websites, Twitter and Facebook. We also identify a certain number of important structural features. A stochastic block model of the hyperlinks structure shows the systematic rejection of counter-informational press in a separate cluster which hardly receives any attention from the mainstream media. Counter-informational sub-spaces are also peripheral on the consumption side. We measure their respective audiences on Twitter and Facebook and do not observe a large discrepancy between both social networks, with counter-information space, far right and far left media gathering limited audiences. Finally, we also measure the ideological distribution of news stories using Twitter data, which also suggests that the French media landscape is quite balanced. We therefore conclude that the French media ecosystem does not suffer from the same level of polarization as the US media ecosystem. The comparison with the American situation also allows us to consolidate a result from studies on disinformation: the polarization of the journalistic space and the circulation of fake news are phenomena that only become more widespread when dominant and influential actors in the political or journalistic space spread topics and dubious content originally circulating in the fringe of the information space.

cs.SI↗

arXiv2021

Your most telling friends: Propagating latent ideological features on Twitter using neighborhood coherence

Multidimensional scaling in networks allows for the discovery of latent information about their structure by embedding nodes in some feature space. Ideological scaling for users in social networks such as Twitter is an example, but similar settings can include diverse applications in other networks and even media platforms or e-commerce. A growing literature of ideology scaling methods in social networks restricts the scaling procedure to nodes that provide interpretability of the feature space: on Twitter, it is common to consider the sub-network of parliamentarians and their followers. This allows to interpret inferred latent features as indices for ideology-related concepts inspecting the position of members of parliament. While effective in inferring meaningful features, this is generally restrained to these sub-networks, limiting interesting applications such as country-wide measurement of polarization and its evolution. We propose two methods to propagate ideological features beyond these sub-networks: one based on homophily (linked users have similar ideology), and the other on structural similarity (nodes with similar neighborhoods have similar ideologies). In our methods, we leverage the concept of neighborhood ideological coherence as a parameter for propagation. Using Twitter data, we produce an ideological scaling for 370K users, and analyze the two families of propagation methods on a population of 6.5M users. We find that, when coherence is considered, the ideology of a user is better estimated from those with similar neighborhoods, than from their immediate neighbors.

cs.SI↗

arXiv2021

Domain-topic models with chained dimensions: charting an emergent domain of a major oncology conference

This paper presents a contribution to the study of bibliographic corpora in the context of science mapping. Starting from a graph representation of documents and their textual dimension, we observe that stochastic block models (SBMs) can provide a simultaneous clustering of documents and words that we call a domain-topic model. Previous work by (Gerlach et al., 2018) investigated the resulting topics, or word clusters, while ours focuses on the study of the document clusters, which we call domains. To enable the synthetic description and interactive navigation of domains, we introduce measures and interfaces relating both types of clusters, which reflect the structure of the graph and the model. We then present a procedure that, starting from the document clusters, extends the block model to also cluster arbitrary metadata attributes of the documents. We call this procedure a domain-chained model, and our previous measures and interfaces can be directly transposed to read the metadata clusters. We provide an example application to a corpus that is relevant to current STS research, and an interesting case for our approach: the 1995-2017 collection of abstracts presented at ASCO, the main annual oncology research conference. Through a sequence of domain-topic and domain-chained models, we identify and describe a particular group of domains in ASCO that have notably grown through the last decades, and which we relate to the establishment of "oncopolicy" as a major concern in oncology.

cs.DL↗

arXiv2015

Detecting global bridges in networks

The identification of nodes occupying important positions in a network structure is crucial for the understanding of the associated real-world system. Usually, betweenness centrality is used to evaluate a node capacity to connect different graph regions. However, we argue here that this measure is not adapted for that task, as it gives equal weight to "local" centers (i.e. nodes of high degree central to a single region) and to "global" bridges, which connect different communities. This distinction is important as the roles of such nodes are different in terms of the local and global organisation of the network structure. In this paper we propose a decomposition of betweenness centrality into two terms, one highlighting the local contributions and the other the global ones. We call the latter bridgeness centrality and show that it is capable to specifically spot out global bridges. In addition, we introduce an effective algorithmic implementation of this measure and demonstrate its capability to identify global bridges in air transportation and scientific collaboration networks.

cs.SI↗

arXiv2013

Citation impacts revisited: how novel impact measures reflect interdisciplinarity and structural change at the local and global level

Citation networks have fed numerous works in scientific evaluation, science mapping (and more recently large-scale network studies) for decades. The variety of citation behavior across scientific fields is both a research topic in sociology of science, and a problem in scientific evaluation. Normalization, tantamount to a particular weighting of links in the citation network, is necessary for allowing across-field comparisons of citation scores and interdisciplinary studies. In addition to classical normalization which drastically reduces all variability factors altogether, two tracks of research have emerged in the recent years. One is the revival of iterative "influence measures". The second is the "citing-side" normalization, whose only purpose is to control for the main factor of variability, the inequality in citing propensity, letting other aspects play: knowledge export/imports and growth. When all variables are defined at the same field-level, two propositions are established: (a) the gross impact measure identifies with the product of relative growth rate, gross balance of citation exchanges, and relative number of references (b) the normalized impact identifies with the product of relative growth rate and normalized balance. At the science level, the variance of growth rate over domains is a proxy for change in the system, and the variance of balance a measure of inter-disciplinary dependences. This opens a new perspective, where the resulting variance of normalized impact, and a related measure, the sum of these variances proposed as a Change-Exchange Indicator, summarize important aspects of science structure and dynamism. Results based on a decade's data are discussed. The behavior of normalized impact according to scale changes is also briefly discussed.

cs.DL↗

arXiv2013

Multi-Level Modeling of Quotation Families Morphogenesis

This paper investigates cultural dynamics in social media by examining the proliferation and diversification of clearly-cut pieces of content: quoted texts. In line with the pioneering work of Leskovec et al. and Simmons et al. on memes dynamics we investigate in deep the transformations that quotations published online undergo during their diffusion. We deliberately put aside the structure of the social network as well as the dynamical patterns pertaining to the diffusion process to focus on the way quotations are changed, how often they are modified and how these changes shape more or less diverse families and sub-families of quotations. Following a biological metaphor, we try to understand in which way mutations can transform quotations at different scales and how mutation rates depend on various properties of the quotations.

cs.CY↗

arXiv2012

Generating constrained random graphs using multiple edge switches

The generation of random graphs using edge swaps provides a reliable method to draw uniformly random samples of sets of graphs respecting some simple constraints, e.g. degree distributions. However, in general, it is not necessarily possible to access all graphs obeying some given con- straints through a classical switching procedure calling on pairs of edges. We therefore propose to get round this issue by generalizing this classical approach through the use of higher-order edge switches. This method, which we denote by "k-edge switching", makes it possible to progres- sively improve the covered portion of a set of constrained graphs, thereby providing an increasing, asymptotically certain confidence on the statistical representativeness of the obtained sample.

cs.SI↗

arXiv2010

Precursors and Laggards: An Analysis of Semantic Temporal Relationships on a Blog Network

We explore the hypothesis that it is possible to obtain information about the dynamics of a blog network by analysing the temporal relationships between blogs at a semantic level, and that this type of analysis adds to the knowledge that can be extracted by studying the network only at the structural level of URL links. We present an algorithm to automatically detect fine-grained discussion topics, characterized by n-grams and time intervals. We then propose a probabilistic model to estimate the temporal relationships that blogs have with one another. We define the precursor score of blog A in relation to blog B as the probability that A enters a new topic before B, discounting the effect created by asymmetric posting rates. Network-level metrics of precursor and laggard behavior are derived from these dyadic precursor score estimations. This model is used to analyze a network of French political blogs. The scores are compared to traditional link degree metrics. We obtain insights into the dynamics of topic participation on this network, as well as the relationship between precursor/laggard and linking behaviors. We validate and analyze results with the help of an expert on the French blogosphere. Finally, we propose possible applications to the improvement of search engine ranking algorithms.

cs.SI↗

arXiv2010

The Reconstruction of Science Phylogeny

We are facing a real challenge when coping with the continuous acceleration of scientific production and the increasingly changing nature of science. In this article, we extend the classical framework of co-word analysis to the study of scientific landscape evolution. Capitalizing on formerly introduced science mapping methods with overlapping clustering, we propose methods to reconstruct phylogenetic networks from successive science maps, and give insight into the various dynamics of scientific domains. Two indexes - the pseudo-inclusion and the empirical quality - are introduced to qualify scientific fields and are used for reconstruction validation purpose. Phylogenetic dynamics appear to be strongly correlated to these two indexes, and to a weaker extent, to a third one previously introduced (density index). These results suggest that there exist regular patterns in the "life cycle" of scientific fields. The reconstruction of science phylogeny should improve our global understanding of science evolution and pave the way toward the development of innovative tools for our daily interactions with its productions. Over the long run, these methods should lead quantitative epistemology up to the point to corroborate or falsify theoretical models of science evolution based on large-scale phylogeny reconstruction from databases of scientific literature.

physics.soc-ph↗

arXiv2010

Academic team formation as evolving hypergraphs

This paper quantitatively explores the social and socio-semantic patterns of constitution of academic collaboration teams. To this end, we broadly underline two critical features of social networks of knowledge-based collaboration: first, they essentially consist of group-level interactions which call for team-centered approaches. Formally, this induces the use of hypergraphs and n-adic interactions, rather than traditional dyadic frameworks of interaction such as graphs, binding only pairs of agents. Second, we advocate the joint consideration of structural and semantic features, as collaborations are allegedly constrained by both of them. Considering these provisions, we propose a framework which principally enables us to empirically test a series of hypotheses related to academic team formation patterns. In particular, we exhibit and characterize the influence of an implicit group structure driving recurrent team formation processes. On the whole, innovative production does not appear to be correlated with more original teams, while a polarization appears between groups composed of experts only or non-experts only, altogether corresponding to collectives with a high rate of repeated interactions.

physics.soc-ph↗

arXiv2009

Socio-semantic dynamics in a blog network

The blogosphere can be construed as a knowledge network made of bloggers who are interacting through a social network to share, exchange or produce information. We claim that the social and semantic dimensions are essentially co-determined and propose to investigate the co-evolutionary dynamics of the blogosphere by examining two intertwined issues: First, how does knowledge distribution drive new interactions and thus influence the social network topology? Second, which role structural network properties play in the information circulation in the system? We adopt an empirical standpoint by analyzing the semantic and social activity of a portion of the US political blogosphere, monitored on a period of four months.

cs.CY↗

arXiv2009

French Roadmap for complex Systems 2008-2009

This second issue of the French Complex Systems Roadmap is the outcome of the Entretiens de Cargese 2008, an interdisciplinary brainstorming session organized over one week in 2008, jointly by RNSC, ISC-PIF and IXXI. It capitalizes on the first roadmap and gathers contributions of more than 70 scientists from major French institutions. The aim of this roadmap is to foster the coordination of the complex systems community on focused topics and questions, as well as to present contributions and challenges in the complex systems sciences and complexity science to the public, political and industrial spheres.

nlin.AO↗