Curriculum learning is often employed in deep reinforcement learning to let the agent progress more quickly towards better behaviors. Numerical methods for curriculum learning in the literature provides only initial heuristic solutions, with little to no guarantee on their quality. We define a new gray-box function that, including a suitable scheduling problem, can be effectively used to reformulate the curriculum learning problem. We propose different efficient numerical methods to address this gray-box reformulation. Preliminary numerical results on a bench- mark task in the curriculum learning literature show the viability of the proposed approach.

A Gray-Box Approach for Curriculum Learning / Foglino, Francesco; Leonetti, Matteo; Sagratella, Simone; Seccia, Ruggiero. - (2019), pp. 720-729. - ADVANCES IN INTELLIGENT SYSTEMS AND COMPUTING. [10.1007/978-3-030-21803-4_72].

A Gray-Box Approach for Curriculum Learning

Leonetti, Matteo;Sagratella, Simone
;
Seccia, Ruggiero
2019

Abstract

Curriculum learning is often employed in deep reinforcement learning to let the agent progress more quickly towards better behaviors. Numerical methods for curriculum learning in the literature provides only initial heuristic solutions, with little to no guarantee on their quality. We define a new gray-box function that, including a suitable scheduling problem, can be effectively used to reformulate the curriculum learning problem. We propose different efficient numerical methods to address this gray-box reformulation. Preliminary numerical results on a bench- mark task in the curriculum learning literature show the viability of the proposed approach.
2019
Optimization of Complex Systems: Theory, Models, Algorithms and Applications
978-3-030-21802-7
978-3-030-21803-4
Curriculum learning; Reinforcement learning; Black-box optimization; Scheduling problem
02 Pubblicazione su volume::02a Capitolo o Articolo
A Gray-Box Approach for Curriculum Learning / Foglino, Francesco; Leonetti, Matteo; Sagratella, Simone; Seccia, Ruggiero. - (2019), pp. 720-729. - ADVANCES IN INTELLIGENT SYSTEMS AND COMPUTING. [10.1007/978-3-030-21803-4_72].
File allegati a questo prodotto
File Dimensione Formato  
Foligno_Postprint_A-Gray-Box_2020.pdf

Open Access dal 21/01/2020

Note: https://link.springer.com/chapter/10.1007/978-3-030-21803-4_72
Tipologia: Documento in Post-print (versione successiva alla peer review e accettata per la pubblicazione)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 396.14 kB
Formato Adobe PDF
396.14 kB Adobe PDF
Foligno_Frontespizio-indice_A-Gray-Box_2020.pdf

solo gestori archivio

Tipologia: Altro materiale allegato
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 253.07 kB
Formato Adobe PDF
253.07 kB Adobe PDF   Contatta l'autore
Foligno_A-Gray-Box_2020.pdf

solo gestori archivio

Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 299.74 kB
Formato Adobe PDF
299.74 kB Adobe PDF   Contatta l'autore

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: https://hdl.handle.net/11573/1288082
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? ND
social impact