ETD system

Electronic theses and dissertations repository

 

Tesi etd-02282019-191815


Thesis type
Tesi di dottorato di ricerca
Author
PEDRELLI, LUCA
email address
lucapedrelli@gmail.com
URN
etd-02282019-191815
Title
Deep Reservoir Computing: A Novel Class of Deep Recurrent Neural Networks
Settore scientifico disciplinare
INF/01
Corso di studi
INFORMATICA
Supervisors
tutor Prof. Micheli, Alessio
tutor Dott. Gallicchio, Claudio
Parole chiave
  • Polyphonic Music Composition
  • Speech Recognition
  • Health Informatics
  • Deep Echo State Networks
  • Reservoir Computing
  • Deep Learning
  • Multivariate Time-series Prediction
  • Deep Recurrent Neural Networks
  • Architectural Design of Recurrent Neural Networks
Data inizio appello
08/03/2019;
Consultabilità
Completa
Riassunto analitico
In this thesis we propose a novel class of deep Recurrent Neural Networks (RNNs) explicitly extending the Reservoir Computing framework to the Deep Learning paradigm. Thereby, we introduce the Deep Echo State Network (DeepESN) model characterized by a hierarchy of randomized recurrent layers.
The introduction of randomized deep RNNs has provided tools to analyze deep recurrent models separately from learning algorithms aspects. The analysis and the experimental assessments conducted on DeepESNs highlighted that layering in deep RNNs is intrinsically able to develop hierarchical, distributed temporal features.
We evaluated our approach on controlled scenarios and challenging real-world tasks.
Overall, DeepESN models allowed us to design extremely efficient deep RNNs that obtained performance competing with state-of-the-art approaches.
File