arXiv subjects

Marc Plantevit

Publications and source records attributed to Marc Plantevit.

14 recordsLinked to original sources

arXiv2026

Differentiable Interval Bottlenecks for Interpretable Anomaly Detection in Numerical Data

Reconstruction-based anomaly detectors are accurate but opaque: a deep autoencoder flags a sample without telling a practitioner which feature ranges made it anomalous. We propose DIFFINT, an autoencoder whose latent bottleneck is structured as a set of soft, axis-aligned interval memberships learned end-to-end directly from raw numerical data, without any discretization or binarization. Each latent unit corresponds to a human-readable hyper-rectangle in feature space; an instance is encoded by how strongly it falls inside each interval relative to the other units, and its reconstruction error is the anomaly score. This keeps the power of differentiable representation learning while exposing an inspectable internal structure. We make the inductive bias precise: a certified reconstruction-error lower bound for points that fall outside every active coordinate of the learned support (with a Lipschitz-enforced decoder), and a graded, empirically verified suppression mechanism for the usual case in which only a few features are abnormal; and we provide a closed-form, label-free importance that ranks each (unit, feature) pair from quantities the model already maintains, turning trained intervals into auditable candidate constraints without ever seeing an anomaly label. On 48 ADBench benchmarks against 22 baselines under a common [-1, 1]-normalized protocol, DIFFINT attains the best mean rank overall on both metrics (4.10 on ROC-AUC, 4.16 on AUPR); among inlier-only detectors it leads its regime clearly, and it is competitive with the strongest contaminated-data detectors (see the stratified and complete-case analyses). It is the only interpretable detector in the statistically-tied leading cluster of seven methods.

cs.LG↗

arXiv2026

ViSAR: Training-Free Adaptive-$k$ Retrieval for Visual Document Question Answering

Document Visual Question Answering (DocVQA) often leverages Retrieval-Augmented Generation (RAG), where late-interaction encoders are commonly used to identify document pages relevant to a user query, before answer generation by a Large Vision-Language Model (LVLM). Existing approaches typically retrieve a fixed top-$k$ number of pages regardless of query complexity, which increases LVLM latency and may degrade answer accuracy. We introduce ViSAR (Visual Semantic Activation Retrieval), a training-free adaptive-$k$ retrieval method for late-interaction visual document retrieval. ViSAR operates directly in the embedding space to construct a query-conditioned page-level similarity matrix that highlights query-relevant semantics and dynamically determines the number of pages to retrieve. Across multiple encoders and LVLMs, ViSAR retrieves compact, query-adapted page sets that reduce RAG latency by up to 58.7\%, while maintaining or improving answer accuracy compared with fixed top-$k$ and adaptive retrieval heuristics. Furthermore, we show that the similarity matrix structure correlates with answer accuracy, suggesting future directions for retrieval quality-aware document understanding.

cs.IR↗

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

RPS: A Generic Reservoir Patterns Sampler

Efficient learning from streaming data is important for modern data analysis due to the continuous and rapid evolution of data streams. Despite significant advancements in stream pattern mining, challenges persist, particularly in managing complex data streams like sequential and weighted itemsets. While reservoir sampling serves as a fundamental method for randomly selecting fixed-size samples from data streams, its application to such complex patterns remains largely unexplored. In this study, we introduce an approach that harnesses a weighted reservoir to facilitate direct pattern sampling from streaming batch data, thus ensuring scalability and efficiency. We present a generic algorithm capable of addressing temporal biases and handling various pattern types, including sequential, weighted, and unweighted itemsets. Through comprehensive experiments conducted on real-world datasets, we evaluate the effectiveness of our method, showcasing its ability to construct accurate incremental online classifiers for sequential data. Our approach not only enables previously unusable online machine learning models for sequential data to achieve accuracy comparable to offline baselines but also represents significant progress in the development of incremental online sequential itemset classifiers.

cs.LG↗

arXiv2024

Scalable Sampling for High Utility Patterns

Discovering valuable insights from data through meaningful associations is a crucial task. However, it becomes challenging when trying to identify representative patterns in quantitative databases, especially with large datasets, as enumeration-based strategies struggle due to the vast search space involved. To tackle this challenge, output space sampling methods have emerged as a promising solution thanks to its ability to discover valuable patterns with reduced computational overhead. However, existing sampling methods often encounter limitations when dealing with large quantitative database, resulting in scalability-related challenges. In this work, we propose a novel high utility pattern sampling algorithm and its on-disk version both designed for large quantitative databases based on two original theorems. Our approach ensures both the interactivity required for user-centered methods and strong statistical guarantees through random sampling. Thanks to our method, users can instantly discover relevant and representative utility pattern, facilitating efficient exploration of the database within seconds. To demonstrate the interest of our approach, we present a compelling use case involving archaeological knowledge graph sub-profiles discovery. Experiments on semantic and none-semantic quantitative databases show that our approach outperforms the state-of-the art methods.

cs.DB↗

arXiv2024

On GNN explanability with activation rules

GNNs are powerful models based on node representation learning that perform particularly well in many machine learning problems related to graphs. The major obstacle to the deployment of GNNs is mostly a problem of societal acceptability and trustworthiness, properties which require making explicit the internal functioning of such models. Here, we propose to mine activation rules in the hidden layers to understand how the GNNs perceive the world. The problem is not to discover activation rules that are individually highly discriminating for an output of the model. Instead, the challenge is to provide a small set of rules that cover all input graphs. To this end, we introduce the subjective activation pattern domain. We define an effective and principled algorithm to enumerate activations rules in each hidden layer. The proposed approach for quantifying the interest of these rules is rooted in information theory and is able to account for background knowledge on the input graph data. The activation rules can then be redescribed thanks to pattern languages involving interpretable features. We show that the activation rules provide insights on the characteristics used by the GNN to classify the graphs. Especially, this allows to identify the hidden features built by the GNN through its different layers. Also, these rules can subsequently be used for explaining GNN decisions. Experiments on both synthetic and real-life datasets show highly competitive performance, with up to 200% improvement in fidelity on explaining graph classification over the SOTA methods.

cs.LG↗

arXiv2024

Peripheral Nervous System Responses to Food Stimuli: Analysis Using Data Science Approaches

In the field of food, as in other fields, the measurement of emotional responses to food and their sensory properties is a major challenge. In the present protocol, we propose a step-by-step procedure that allows a physiological description of odors, aromas, and their hedonic properties. The method rooted in subgroup discovery belongs to the field of data science and especially data mining. It is still little used in the field of food and is based on a descriptive modeling of emotions on the basis of human physiological responses.

q-bio.NC↗

arXiv2023

Mining Java Memory Errors using Subjective Interesting Subgroups with Hierarchical Targets

Software applications, especially Enterprise Resource Planning (ERP) systems, are crucial to the day-to-day operations of many industries. Therefore, it is essential to maintain these systems effectively using tools that can identify, diagnose, and mitigate their incidents. One promising data-driven approach is the Subgroup Discovery (SD) technique, a data mining method that can automatically mine incident datasets and extract discriminant patterns to identify the root causes of issues. However, current SD solutions have limitations in handling complex target concepts with multiple attributes organized hierarchically. To illustrate this scenario, we examine the case of Java out-of-memory incidents among several possible applications. We have a dataset that describes these incidents, including their context and the types of Java objects occupying memory when it reaches saturation, with these types arranged hierarchically. This scenario inspires us to propose a novel Subgroup Discovery approach that can handle complex target concepts with hierarchies. To achieve this, we design a pattern syntax and a quality measure that ensure the identified subgroups are relevant, non-redundant, and resilient to noise. To achieve the desired quality measure, we use the Subjective Interestingness model that incorporates prior knowledge about the data and promotes patterns that are both informative and surprising relative to that knowledge. We apply this framework to investigate out-of-memory errors and demonstrate its usefulness in incident diagnosis. To validate the effectiveness of our approach and the quality of the identified patterns, we present an empirical study. The source code and data used in the evaluation are publicly accessible, ensuring transparency and reproducibility.

cs.IT↗

arXiv2021

Interpretable Summaries of Black Box Incident Triaging with Subgroup Discovery

The need of predictive maintenance comes with an increasing number of incidents reported by monitoring systems and equipment/software users. In the front line, on-call engineers (OCEs) have to quickly assess the degree of severity of an incident and decide which service to contact for corrective actions. To automate these decisions, several predictive models have been proposed, but the most efficient models are opaque (say, black box), strongly limiting their adoption. In this paper, we propose an efficient black box model based on 170K incidents reported to our company over the last 7 years and emphasize on the need of automating triage when incidents are massively reported on thousands of servers running our product, an ERP. Recent developments in eXplainable Artificial Intelligence (XAI) help in providing global explanations to the model, but also, and most importantly, with local explanations for each model prediction/outcome. Sadly, providing a human with an explanation for each outcome is not conceivable when dealing with an important number of daily predictions. To address this problem, we propose an original data-mining method rooted in Subgroup Discovery, a pattern mining technique with the natural ability to group objects that share similar explanations of their black box predictions and provide a description for each group. We evaluate this approach and present our preliminary results which give us good hope towards an effective OCE's adoption. We believe that this approach provides a new way to address the problem of model agnostic outcome explanation.

cs.AI↗

arXiv2020

Sequential recommendation with metric models based on frequent sequences

Modeling user preferences (long-term history) and user dynamics (short-term history) is of greatest importance to build efficient sequential recommender systems. The challenge lies in the successful combination of the whole user's history and his recent actions (sequential dynamics) to provide personalized recommendations. Existing methods capture the sequential dynamics of a user using fixed-order Markov chains (usually first order chains) regardless of the user, which limits both the impact of the past of the user on the recommendation and the ability to adapt its length to the user profile. In this article, we propose to use frequent sequences to identify the most relevant part of the user history for the recommendation. The most salient items are then used in a unified metric model that embeds items based on user preferences and sequential dynamics. Extensive experiments demonstrate that our method outperforms state-of-the-art, especially on sparse datasets. We show that considering sequences of varying lengths improves the recommendations and we also emphasize that these sequences provide explanations on the recommendation.

cs.LG↗

arXiv2019

Mining Subjectively Interesting Attributed Subgraphs

Community detection in graphs, data clustering, and local pattern mining are three mature fields of data mining and machine learning. In recent years, attributed subgraph mining is emerging as a new powerful data mining task in the intersection of these areas. Given a graph and a set of attributes for each vertex, attributed subgraph mining aims to find cohesive subgraphs for which (a subset of) the attribute values has exceptional values in some sense. While research on this task can borrow from the three abovementioned fields, the principled integration of graph and attribute data poses two challenges: the definition of a pattern language that is intuitive and lends itself to efficient search strategies, and the formalization of the interestingness of such patterns. We propose an integrated solution to both of these challenges. The proposed pattern language improves upon prior work in being both highly flexible and intuitive. We show how an effective and principled algorithm can enumerate patterns of this language. The proposed approach for quantifying interestingness of patterns of this language is rooted in information theory, and is able to account for prior knowledge on the data. Prior work typically quantifies interestingness based on the cohesion of the subgraph and for the exceptionality of its attributes separately, combining these in a parametrized trade-off. Instead, in our proposal this trade-off is implicitly handled in a principled, parameter-free manner. Extensive empirical results confirm the proposed pattern syntax is intuitive, and the interestingness measure aligns well with actual subjective interestingness.

cs.SI↗

arXiv2015

Interpreting communities based on the evolution of a dynamic attributed network

Many methods have been proposed to detect communities, not only in plain, but also in attributed, directed or even dynamic complex networks. From the modeling point of view, to be of some utility, the community structure must be characterized relatively to the properties of the studied system. However, most of the existing works focus on the detection of communities, and only very few try to tackle this interpretation problem. Moreover, the existing approaches are limited either by the type of data they handle, or by the nature of the results they output. In this work, we see the interpretation of communities as a problem independent from the detection process, consisting in identifying the most characteristic features of communities. We give a formal definition of this problem and propose a method to solve it. To this aim, we first define a sequence-based representation of networks, combining temporal information, community structure, topological measures, and nodal attributes. We then describe how to identify the most emerging sequential patterns of this dataset, and use them to characterize the communities. We study the performance of our method on artificially generated dynamic attributed networks. We also empirically validate our framework on real-world systems: a DBLP network of scientific collaborations, and a LastFM network of social and musical interactions.

cs.SI↗

arXiv2014

A Method for Characterizing Communities in Dynamic Attributed Complex Networks

Many methods have been proposed to detect communities, not only in plain, but also in attributed, directed or even dynamic complex networks. In its simplest form, a community structure takes the form of a partition of the node set. From the modeling point of view, to be of some utility, this partition must then be characterized relatively to the properties of the studied system. However, if most of the existing works focus on defining methods for the detection of communities, only very few try to tackle this interpretation problem. Moreover, the existing approaches are limited either in the type of data they handle, or by the nature of the results they output. In this work, we propose a method to efficiently support such a characterization task. We first define a sequence-based representation of networks, combining temporal information, topological measures, and nodal attributes. We then describe how to identify the most emerging sequential patterns of this dataset, and use them to characterize the communities. We also show how to detect unusual behavior in a community, and highlight outliers. Finally, as an illustration, we apply our method to a network of scientific collaborations.

cs.SI↗

arXiv2013

Une m\'ethode pour caract\'eriser les communaut\'es des r\'eseaux dynamiques \`a attributs

Many complex systems are modeled through complex networks whose analysis reveals typical topological properties. Amongst those, the community structure is one of the most studied. Many methods are proposed to detect communities, not only in plain, but also in attributed, directed or even dynamic networks. A community structure takes the form of a partition of the node set, which must then be characterized relatively to the properties of the studied system. We propose a method to support such a characterization task. We define a sequence-based representation of networks, combining temporal information, topological measures, and nodal attributes. We then characterize communities using the most representative emerging sequential patterns of its nodes. This also allows detecting unusual behavior in a community. We describe an empirical study of a network of scientific collaborations.---De nombreux syst\`emes complexes sont \'etudi\'es via l'analyse de r\'eseaux dits complexes ayant des propri\'et\'es topologiques typiques. Parmi cellesci, les structures de communaut\'es sont particuli\`erement \'etudi\'ees. De nombreuses m\'ethodes permettent de les d\'etecter, y compris dans des r\'eseaux contenant des attributs nodaux, des liens orient\'es ou \'evoluant dans le temps. La d\'etection prend la forme d'une partition de l'ensemble des noeuds, qu'il faut ensuite caract\'eriser relativement au syst\`eme mod\'elis\'e. Nous travaillons sur l'assistance \`a cette t\^ache de caract\'erisation. Nous proposons une repr\'esentation des r\'eseaux sous la forme de s\'equences de descripteurs de noeuds, qui combinent les informations temporelles, les mesures topologiques, et les valeurs des attributs nodaux. Les communaut\'es sont caract\'eris\'ees au moyen des motifs s\'equentiels \'emergents les plus repr\'esentatifs issus de leurs noeuds. Ceci permet notamment la d\'etection de comportements inhabituels au sein d'une communaut\'e. Nous d\'ecrivons une \'etude empirique sur un r\'eseau de collaboration scientifique.

cs.SI↗