arXiv subjects

Rémy Cazabet

Publications and source records attributed to Rémy Cazabet.

At least 19 recordsLinked to original sources

arXiv2026

C-Unseen: Weak Signal Detection in Dynamic Temporal Knowledge Graphs via LLM Reasoning

Weak signals are early, low-visibility indicators that precede significant changes before those changes become established. Existing detection methods, based on keyword frequency, topic modeling, or untyped graph topology, fail to capture the semantic and relational structure through which such signals manifest. In this paper, we propose C-Unseen, a self-interpretable framework for weak signal detection in Dynamic Temporal Knowledge Graphs (DTKGs). We define a weak signal as a rare, semantically coherent subgraph that proliferates across consecutive TKG snapshots. The framework operates through two modules: a Rare Subgraphs Extractor, in which an LLM identifies subgraphs whose content is in tension with the dominant snapshot narrative via chain-of-thought reasoning, and a Weak Signal Alerter, in which the persistence of these rare subgraphs is tracked across time steps to isolate true weak signals. Experimental results demonstrate that C-Unseen outperforms keyword-, topic-, and graph-based baselines.

cs.AI↗

arXiv2026

Instance Discrimination for Link Prediction

Recently, instance discrimination models have emerged as a major solution for self-supervised learning. Having already demonstrated its effectiveness in the image domain, instance discrimination learning is now proving equally convincing in the graph domain, in particular for node classification. However, fewer contributions have tackled the link prediction task. In this contribution, we propose to adapt existing methods to this context. We first provide a rigorous evaluation of existing self-supervised models in the field of link prediction, showing that the main performance depends on the augmentation process (like in computer vision). We then propose a new structural augmentation based on the community structure that is relevant for link prediction. Our main contribution introduces two new models, L-GRACE and L-BGRL, based on link representations instead of node representations, which improve the performance of the existing methods, especially on unattributed graphs, and we show that they perform on par with the state of the art, both in supervised and self-supervised contexts.

cs.LG↗

arXiv2026

ATOM: AdapTive and OptiMized dynamic temporal knowledge graph construction using LLMs

In today's rapidly expanding data landscape, knowledge extraction from unstructured text is vital for real-time analytics, temporal inference, and dynamic memory frameworks. However, traditional static knowledge graph (KG) construction often overlooks the dynamic and time-sensitive nature of real-world data, limiting adaptability to continuous changes. Moreover, recent zero- or few-shot approaches that avoid domain-specific fine-tuning or reliance on prebuilt ontologies often suffer from instability across multiple runs, as well as incomplete coverage of key facts. To address these challenges, we introduce ATOM (AdapTive and OptiMized), a few-shot and scalable approach that builds and continuously updates Temporal Knowledge Graphs (TKGs) from unstructured texts. ATOM splits input documents into minimal, self-contained "atomic" facts, improving extraction exhaustivity and stability. Then, it constructs atomic TKGs from these facts, employing a dual-time modeling that distinguishes between when information is observed and when it is valid. The resulting atomic TKGs are subsequently merged in parallel. Empirical evaluations demonstrate that ATOM achieves ~18% higher exhaustivity, ~33% better stability, and over ~90% latency reduction compared to baseline methods, demonstrating a strong scalability potential for dynamic TKG construction.

cs.AI↗

arXiv2025

Discovering Communities in Continuous-Time Temporal Networks by Optimizing L-Modularity

Community detection is a fundamental problem in network analysis, with many applications in various fields. Extending community detection to the temporal setting with exact temporal accuracy, as required by real-world dynamic data, necessitates methods specifically adapted to the temporal nature of interactions. We introduce LAGO, a novel method for uncovering dynamic communities by greedy optimization of Longitudinal Modularity, a specific adaptation of Modularity for continuous-time networks. Unlike prior approaches that rely on time discretization or assume rigid community evolution, LAGO captures the precise moments when nodes enter and exit communities. We evaluate LAGO on synthetic benchmarks and real-world datasets, demonstrating its ability to efficiently uncover temporally and topologically coherent communities.

cs.SI↗

arXiv2025

Cryptocurrency Network Analysis

Cryptocurrency network analysis consists of applying the tools and methods of social network analysis to transactional data issued from cryptocurrencies. The main difference with most online social networks is that users do not exchange textual content but instead value -- in systems designed mainly as cryptocurrency, such as Bitcoin -- or digital items and services in more permissive systems based on smart contracts such as Ethereum.

cs.SI↗

arXiv2025

SEANN: A Domain-Informed Neural Network for Epidemiological Insights

In epidemiology, traditional statistical methods such as logistic regression, linear regression, and other parametric models are commonly employed to investigate associations between predictors and health outcomes. However, non-parametric machine learning techniques, such as deep neural networks (DNNs), coupled with explainable AI (XAI) tools, offer new opportunities for this task. Despite their potential, these methods face challenges due to the limited availability of high-quality, high-quantity data in this field. To address these challenges, we introduce SEANN, a novel approach for informed DNNs that leverages a prevalent form of domain-specific knowledge: Pooled Effect Sizes (PES). PESs are commonly found in published Meta-Analysis studies, in different forms, and represent a quantitative form of a scientific consensus. By direct integration within the learning procedure using a custom loss, we experimentally demonstrate significant improvements in the generalizability of predictive performances and the scientific plausibility of extracted relationships compared to a domain-knowledge agnostic neural network in a scarce and noisy data setting.

cs.LG↗

arXiv2024

Inside Alameda Research: A Multi-Token Network Analysis

We analyze the token transfer network on Ethereum, focusing on accounts associated with Alameda Research, a cryptocurrency trading firm implicated in the misuse of FTX customer funds. Using a multi-token network representation, we examine node centralities and the network backbone to identify critical accounts, tokens, and activity groups. The temporal evolution of Alameda accounts reveals shifts in token accumulation and distribution patterns leading up to its bankruptcy in November 2022. Through network analysis, our work offers insights into the activities and dynamics that shape the DeFi ecosystem.

cs.SI↗

arXiv2024

iText2KG: Incremental Knowledge Graphs Construction Using Large Language Models

Most available data is unstructured, making it challenging to access valuable information. Automatically building Knowledge Graphs (KGs) is crucial for structuring data and making it accessible, allowing users to search for information effectively. KGs also facilitate insights, inference, and reasoning. Traditional NLP methods, such as named entity recognition and relation extraction, are key in information retrieval but face limitations, including the use of predefined entity types and the need for supervised learning. Current research leverages large language models' capabilities, such as zero- or few-shot learning. However, unresolved and semantically duplicated entities and relations still pose challenges, leading to inconsistent graphs and requiring extensive post-processing. Additionally, most approaches are topic-dependent. In this paper, we propose iText2KG, a method for incremental, topic-independent KG construction without post-processing. This plug-and-play, zero-shot method is applicable across a wide range of KG construction scenarios and comprises four modules: Document Distiller, Incremental Entity Extractor, Incremental Relation Extractor, and Graph Integrator and Visualization. Our method demonstrates superior performance compared to baseline methods across three scenarios: converting scientific papers to graphs, websites to graphs, and CVs to graphs.

cs.AI↗

arXiv2024

ORBITAAL: A Temporal Graph Dataset of Bitcoin Entity-Entity Transactions

Research on Bitcoin (BTC) transactions is a matter of interest for both economic and network science fields. Although this cryptocurrency is based on a decentralized system, making transaction details freely accessible, making raw blockchain data analyzable is not straightforward due to the Bitcoin protocol specificity and data richness. To address the need for an accessible dataset, we present ORBITAAL, the first comprehensive dataset based on temporal graph formalism. The dataset covers all Bitcoin transactions from January 2009 to January 2021. ORBITAAL provides temporal graph representations of entity-entity transaction networks, snapshots and stream graph. Each transaction value is given in Bitcoin and US dollar regarding daily-based conversion rate. This dataset also provides details on entities such as their global BTC balance and associated public addresses.

cs.SI↗

arXiv2024

Decoding Decentralized Finance Transactions through Ego Network Motif Mining

Decentralized Finance (DeFi) is increasingly studied and adopted for its potential to provide accessible and transparent financial services. Analyzing how investors use DeFi is important for reaching a better understanding of their usage and for regulation purposes. However, analyzing DeFi transactions is challenging due to often incomplete or inaccurate labeled data. This paper presents a method to extract ego network motifs from the token transfer network, capturing the transfer of tokens between users and smart contracts. Our results demonstrate that smart contract methods performing specific DeFi operations can be efficiently identified by analyzing these motifs while providing insights into account activities.

cs.SI↗

arXiv2024

Y Social: an LLM-powered Social Media Digital Twin

In this paper we introduce Y, a new-generation digital twin designed to replicate an online social media platform. Digital twins are virtual replicas of physical systems that allow for advanced analyses and experimentation. In the case of social media, a digital twin such as Y provides a powerful tool for researchers to simulate and understand complex online interactions. {\tt Y} leverages state-of-the-art Large Language Models (LLMs) to replicate sophisticated agent behaviors, enabling accurate simulations of user interactions, content dissemination, and network dynamics. By integrating these aspects, Y offers valuable insights into user engagement, information spread, and the impact of platform policies. Moreover, the integration of LLMs allows Y to generate nuanced textual content and predict user responses, facilitating the study of emergent phenomena in online environments. To better characterize the proposed digital twin, in this paper we describe the rationale behind its implementation, provide examples of the analyses that can be performed on the data it enables to be generated, and discuss its relevance for multidisciplinary research.

cs.AI↗

arXiv2024

Redefining Event Types and Group Evolution in Temporal Data

Groups -- such as clusters of points or communities of nodes -- are fundamental when addressing various data mining tasks. In temporal data, the predominant approach for characterizing group evolution has been through the identification of ``events". However, the events usually described in the literature, e.g., shrinks/growths, splits/merges, are often arbitrarily defined, creating a gap between such theoretical/predefined types and real-data group observations. Moving beyond existing taxonomies, we think of events as ``archetypes" characterized by a unique combination of quantitative dimensions that we call ``facets". Group dynamics are defined by their position within the facet space, where archetypal events occupy extremities. Thus, rather than enforcing strict event types, our approach can allow for hybrid descriptions of dynamics involving group proximity to multiple archetypes. We apply our framework to evolving groups from several face-to-face interaction datasets, showing it enables richer, more reliable characterization of group dynamics with respect to state-of-the-art methods, especially when the groups are subject to complex relationships. Our approach also offers intuitive solutions to common tasks related to dynamic group analysis, such as choosing an appropriate aggregation scale, quantifying partition stability, and evaluating event quality.

cs.LG↗

arXiv2022

Pattern Analysis of Money Flow in the Bitcoin Blockchain

Bitcoin is the first and highest valued cryptocurrency that stores transactions in a publicly distributed ledger called the blockchain. Understanding the activity and behavior of Bitcoin actors is a crucial research topic as they are pseudonymous in the transaction network. In this article, we propose a method based on taint analysis to extract taint flows --dynamic networks representing the sequence of Bitcoins transferred from an initial source to other actors until dissolution. Then, we apply graph embedding methods to characterize taint flows. We evaluate our embedding method with taint flows from top mining pools and show that it can classify mining pools with high accuracy. We also found that taint flows from the same period show high similarity. Our work proves that tracing the money flows can be a promising approach to classifying source actors and characterizing different money flow patterns

cs.SI↗

arXiv2022

Graph space: using both geometric and probabilistic structure to evaluate statistical graph models

Statistical graph models aim at modeling graphs as random realization among a set of possible graphs. One issue is to evaluate whether or not a graph is likely to have been generated by one particular model. In this paper we introduce the edit distance expected value (EDEV) and compare it with other methods such as entropy and distance to the barycenter. We show that contrary to them, EDEV is able to distinguish between graphs that have a typical structure with respect to a model, and those that do not. Finally we introduce a statistical hypothesis testing methodology based on this distance to evaluate the relevance of a candidate model with respect to an observed graph.

cs.SI↗

arXiv2021

Quantitative Evaluation of Snapshot Graphs for the Analysis of Temporal Networks

One of the most common approaches to the analysis of dynamic networks is through time-window aggregation. The resulting representation is a sequence of static networks, i.e. the snapshot graph. Despite this representation being widely used in the literature, a general framework to evaluate the soundness of snapshot graphs is still missing. In this article, we propose two scores to quantify conflicting objectives: Stability measures how much stable the sequence of snapshots is, while Fidelity measures the loss of information compared to the original data. We also develop a technique of targeted filtering of the links, to simplify the original temporal network. Our framework is tested on datasets of proximity and face-to-face interactions.

cs.SI↗

arXiv2021

Δ-Conformity: Multi-scale Node Assortativity in Feature-rich Stream Graphs

Heterogeneity is a key aspect of complex networks, often emerging by looking at the distribution of node properties, from the milestone observations on the degree to the recent developments in mixing pattern estimation. Mixing patterns, in particular, refer to nodes' connectivity preferences with respect to an attribute label. Social networks are mostly characterized by assortative/homophilic behaviour, where nodes are more likely to be connected with similar ones. Recently, assortative mixing is increasingly measured in a multi-scale fashion to overcome well known limitations of classic scores. Such multi-scale strategies can capture heterogeneous behaviors among node homophily, but they ignore an important, often available, addendum in real-world systems: the time when edges are present and the time-varying paths they form accordingly. Hence, temporal homophily is still little understood in complex networks. In this work we aim to cover this gap by introducing the Δ-Conformity measure, a multiscale, path-aware, node homophily estimator within the new framework of feature-rich stream graphs. A rich experimental section analyzes Δ-Conformity trends over time, spanning the analysis from real-life social interaction networks to a specific case-study about the Bitcoin Transaction Network.

cs.SI↗

arXiv2021

Graph model selection by edge probability sequential inference

Graphs are widely used for describing systems made up of many interacting components and for understanding the structure of their interactions. Various statistical models exist, which describe this structure as the result of a combination of constraints and randomness. %Model selection techniques need to automatically identify the best model, and the best set of parameters for a given graph. To do so, most authors rely on the minimum description length paradigm, and apply it to graphs by considering the entropy of probability distributions defined on graph ensembles. In this paper, we introduce edge probability sequential inference, a new approach to perform model selection, which relies on probability distributions on edge ensembles. From a theoretical point of view, we show that this methodology provides a more consistent ground for statistical inference with respect to existing techniques, due to the fact that it relies on multiple realizations of the random variable. It also provides better guarantees against overfitting, by making it possible to lower the number of parameters of the model below the number of observations. Experimentally, we illustrate the benefits of this methodology in two situations: to infer the partition of a stochastic blockmodel, and to identify the most relevant model for a given graph between the stochastic blockmodel and the configuration model.

stat.ME↗

arXiv2021

Edge based stochastic block model statistical inference

Community detection in graphs often relies on ad hoc algorithms with no clear specification about the node partition they define as the best, which leads to uninterpretable communities. Stochastic block models (SBM) offer a framework to rigorously define communities, and to detect them using statistical inference method to distinguish structure from random fluctuations. In this paper, we introduce an alternative definition of SBM based on edge sampling. We derive from this definition a quality function to statistically infer the node partition used to generate a given graph. We then test it on synthetic graphs, and on the zachary karate club network.

cs.SI↗