Kirell Maël Benzi

Official source

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Unités associées

Chargement

Cours enseignés par cette personne

Chargement

Domaines de recherche associés

Chargement

Publications associées

Chargement

Personnes menant des recherches similaires

Chargement

Anomaly detection in the dynamics of web and social networks

Kirell Maël Benzi, Volodymyr Miz, Benjamin Ricaud, Pierre Vandergheynst

In this work, we propose a new, fast and scalable method for anomaly detection in large time-evolving graphs. It may be a static graph with dynamic node attributes (e.g. time-series), or a graph evolving in time, such as a temporal network. We define an anomaly as a localized increase in temporal activity in a cluster of nodes. The algorithm is unsupervised. It is able to detect and track anomalous activity in a dynamic network despite the noise from multiple interfering sources. We use the Hopfield network model of memory to combine the graph and time information. We show that anomalies can be spotted with good precision using a memory network. The presented approach is scalable and we provide a distributed implementation of the algorithm. To demonstrate its efficiency, we apply it to two datasets: Enron Email dataset and Wikipedia page views. We show that the anomalous spikes are triggered by the real-world events that impact the network dynamics. Besides, the structure of the clusters and the analysis of the time evolution associated with the detected events reveals interesting facts on how humans interact, exchange and search for information, opening the door to new quantitative studies on collective and social behavior on large and dynamic datasets.

2019

Chargement

, , ,

ASSOC COMPUTING MACHINERY2019

Chargement

FMA: A Dataset For Music Analysis

Kirell Maël Benzi, Xavier Bresson, Michaël Defferrard, Pierre Vandergheynst

We introduce the Free Music Archive (FMA), an open and easily accessible dataset which can be used to evaluate several tasks in music information retrieval (MIR), a field concerned with browsing, searching, and organizing large music collections. The community's growing interest in feature and end-to-end learning is however restrained by the limited availability of large audio datasets. By releasing the FMA, we hope to foster research which will improve the state-of-the-art and hopefully surpass the performance ceiling observed in e.g. genre recognition (MGR). The data is made of 106,574 tracks, 16,341 artists, 14,854 albums, arranged in a hierarchical taxonomy of 161 genres, for a total of 343 days of audio and 917 GiB, all under permissive Creative Commons licenses. It features metadata like song title, album, artist and genres; user data like play counts, favorites, and comments; free-form text like description, biography, and tags; together with full-length, high-quality audio, and some pre-computed features. We propose a train/validation/test split and three subsets: a genre-balanced set of 8,000 tracks from 8 major genres, a genre-unbalanced set of 25,000 tracks from 16 genres, and a 98 GiB version with clips trimmed to 30s. This paper describes the dataset and how it was created, proposes some tasks like music classification and annotation or recommendation, and evaluates some baselines for MGR. Code, data, and usage examples are available at https://github.com/mdeff/fma.

2017

Chargement