Kernel methods have been used by many Machine Learning paradigms, achieving state-of-the-art performances in many Language Learning tasks. One drawback of expressive kernel functions, such as Sequence or Tree kernels, is the time and space complexity required both in learning and classification. In this paper, the Nyström methodology is studied as a viable solution to face these scalability issues. By mapping data in low-dimensional spaces as kernel space approximations, the proposed methodology positively impacts on scalability through compact linear representation of highly structured data. Computation can be also distributed on several machines by adopting the so-called Ensemble Nyström Method. Experimental results show that an accuracy comparable with state-of-the-art kernel-based methods can be obtained by reducing of orders of magnitude the required operations and enabling the adoption of datasets containing more than one million examples.

Croce, D., & Basili, R. (2016). Large-scale Kernel-based language learning through the ensemble Nyström methods. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp.100-112). Springer Verlag [10.1007/978-3-319-30671-1_8].

Large-scale Kernel-based language learning through the ensemble Nyström methods

CROCE, DANILO;BASILI, ROBERTO
2016-01-01

Abstract

Kernel methods have been used by many Machine Learning paradigms, achieving state-of-the-art performances in many Language Learning tasks. One drawback of expressive kernel functions, such as Sequence or Tree kernels, is the time and space complexity required both in learning and classification. In this paper, the Nyström methodology is studied as a viable solution to face these scalability issues. By mapping data in low-dimensional spaces as kernel space approximations, the proposed methodology positively impacts on scalability through compact linear representation of highly structured data. Computation can be also distributed on several machines by adopting the so-called Ensemble Nyström Method. Experimental results show that an accuracy comparable with state-of-the-art kernel-based methods can be obtained by reducing of orders of magnitude the required operations and enabling the adoption of datasets containing more than one million examples.
38th European Conference on Information Retrieval Research, ECIR 2016
ita
2016
Elsevier
Rilevanza internazionale
Settore ING-INF/05 - Sistemi di Elaborazione delle Informazioni
English
Computer Science (all); Theoretical Computer Science
http://springerlink.com/content/0302-9743/copyright/2005/
Intervento a convegno
Croce, D., & Basili, R. (2016). Large-scale Kernel-based language learning through the ensemble Nyström methods. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp.100-112). Springer Verlag [10.1007/978-3-319-30671-1_8].
Croce, D; Basili, R
File in questo prodotto:
File Dimensione Formato  
ECIR2016_v1.3.pdf

non disponibili

Licenza: Copyright dell'editore
Dimensione 338.64 kB
Formato Adobe PDF
338.64 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/2108/157907
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? ND
social impact