Detecting hazardous events from sequential data with multilayer architectures

Karol Kurach, Krzysztof Pawlowski

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Multivariate time series data play an important role in many domains, including real-time monitoring systems. In this paper, we focus on multilayer neural architectures that are capable of learning high level representations from raw data. This includes our previous solution based on Recurrent Neural Networks with Long Short-Term Memory (LSTM) cells. We build upon this work and present improved methods that aim to achieve higher prediction quality and better generalization to other similar tasks. We apply new deep neural architectures, minimize feature engineering and explore different ways of model selection. In particular, our focus on architectures includes networks with attention mechanism and convolutional networks. We tackle overfitting challenges in a presence of concept drift.

Original languageEnglish (US)
Title of host publicationConcurrency, Specification and Programming - 24th InternationalWorkshop, CS and P 2015, Proceedings
EditorsZbigniew Suraj, Ludwik Czaja, Ludwik Czaja
PublisherCEUR-WS
Pages1-10
Number of pages10
ISBN (Electronic)9788379961818
StatePublished - 2015
Externally publishedYes
Event24th International Workshop on Concurrency, Specification and Programming, CS and P 2015 - Rzeszow, Poland
Duration: Sep 28 2015Sep 30 2015

Publication series

NameCEUR Workshop Proceedings
Volume1492
ISSN (Print)1613-0073

Conference

Conference24th International Workshop on Concurrency, Specification and Programming, CS and P 2015
Country/TerritoryPoland
CityRzeszow
Period9/28/159/30/15

Keywords

  • Deep learning
  • Ensemble methods
  • Recurrent neural networks
  • Time series forecasting

ASJC Scopus subject areas

  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Detecting hazardous events from sequential data with multilayer architectures'. Together they form a unique fingerprint.

Cite this