Título:
Timeless foundations : Exploring DC-VAEs as foundation models for time series analysis.
Otros títulos:
Coordinador:
Director:
Compilador:
Autor:
García González, Gastón
Casas, Pedro
Martínez, Emilio
Fernández, Alicia
Casas, Pedro
Martínez, Emilio
Fernández, Alicia
Tutor:
Tipo de documento:
Preprint
Editor:
Palabras clave:
Time-Series Data
Generative AI
Anomaly Detection
VAE
Foundation Models for Time-Series
Generative AI
Anomaly Detection
VAE
Foundation Models for Time-Series
Descriptores:
Año de publicación:
2024
Contenido:
Resumen:
We investigate a novel approach to time-series modeling, inspired by the successes of large pre-trained foundation models. We introduce FAE (Foundation Auto-Encoders), a foundation generative-AI model for anomaly detection in time-series data, based on Variational Auto-Encoders (VAEs). By Foundation Model (FM), we mean a model pre-trained on massive amounts of time-series data which can learn complex temporal patterns useful for accurate modeling and forecasting on previously unseen datasets. FAE leverages VAEs and Dilated Convolutional Neural Networks (DCNNs) to build a generic model for time-series modeling, which could eventually perform properly in out-of-the-box, zero-shot anomaly detection applications. We introduce the main concepts and ideas of this FM for time-series (TSFM), and present some preliminary results in a multi-dimensional mobile network monitoring dataset. We also present example results applying novel TSFMs to this dataset, both in a zero-shot manner and relying on fine-tuning, and show how complex it is in the practice to achieve accurate results.
Descripción:
metadata.articulos.dc.description.uri:
Editorial:
EN:
Financiadores:
Este trabajo ha sido financiado parcialmente por el Austrian FFG ICT-of-the-Future project DynAISEC – Adaptive AI/ML for Dynamic Cybersecurity Systems – project ID 887504, the Austrian Science Fund, FWF project GRAPHS4SEC – Graph Neural Networks for Robust AI/ML-driven Network Security Applications – grant number I-6653, y por el proyecto uruguayo CSIC Generalization and Domain Adap- tation in Time-Series Anomaly Detection, project ID CSIC- I+D-22520220100371UD.
Citación:
García González, G., Casas, P., Martínez, E. y otros. Timeless foundations : Exploring DC-VAEs as foundation models for time series analysis [Preprint]. Publicado en: 2024 8th Network Traffic Measurement and Analysis Conference (TMA), Dresden, Germany, 21-24 may. 2024, pp. 1-4.
Citación:
metadata.articulos.cc.license.name:
ISBN:
e-ISBN:
ISSN:
ISMN:
Otros identificadores:
Cobertura geográfica:
Cobertura temporal:
Departamento académico:
Procesamiento de Señales
Grupo de investigación:
Tratamiento de Imagenes
Licencia:
Licencia Creative Commons Atribución - No Comercial - Sin Derivadas (CC - By-NC-ND 4.0)
Colecciones:
| Ficheros | Descripción | Tamaño | Formato | ||
|---|---|---|---|---|---|
| GCMF24b.pdf | Preprint | 1.27 MB | Adobe PDF |