Learning to Look in Different Environments: An Active-Vision Model which Learns and Readapts Visual Routines (Articolo in rivista)

Type

• Prodotto della ricerca (Classe)
• Articolo in rivista (Classe)

Label

• Learning to Look in Different Environments: An Active-Vision Model which Learns and Readapts Visual Routines (Articolo in rivista) (literal)

Anno

• 2010-01-01T00:00:00+01:00 (literal)

Alternative label

• Ognibene D.; Pezzulo G.; Baldassarre G. (2010)
  Learning to Look in Different Environments: An Active-Vision Model which Learns and Readapts Visual Routines
  in Lecture notes in computer science
  (literal)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#autori

• Ognibene D.; Pezzulo G.; Baldassarre G. (literal)

Pagina inizio

• 199 (literal)

Pagina fine

• 210 (literal)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#numeroVolume

• 6226/ (literal)

Rivista

• Lecture notes in computer science (Rivista)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#note

• In: Lecture Notes in Computer Science, vol. 6226/2010 pp. 199 - 210. From Animals to Animats 11. 11th International Conference on Simulation of Adaptive Behavior, SAB 2010, Paris - Clos Lucé, France, August 25-28, 2010. Proceedings. Doncieux, S.; Girard, B.; Guillot, A.; Hallam, J.; Meyer, J.-A.; Mouret, J.-B. (Eds.) (eds.). Springer, 2010. (literal)

Note

• ISI Web of Science (WOS) (literal)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#affiliazioni

• CNR-ISTC, Roma, CNR-ILC, Pisa (literal)

Titolo

• Learning to Look in Different Environments: An Active-Vision Model which Learns and Readapts Visual Routines (literal)

Abstract

• One of the main claims of the active vision framework is that nding data on the basis of task requirements is more e?cient than reconstructing the whole scene by performing a complete visual scan. To be successful, this approach requires that agents learn visual routines to direct overt attention to locations with the information needed to accomplish the task. In ecological conditions, learning such visual routines is di?cult due to the partial observability of the world, the changes in the environment, and the fact that learning signals might be indirect. This paper uses a reinforcement-learning actor-critic model to study how visual routines can be formed, and then adapted when the environment changes, in a system endowed with a controllable gaze and reaching capabilities. The tests of the model show that: (a) the autonomouslydeveloped visual routines are strongly dependent on the task and the statistical properties of the environment; (b) when the statistics of the environment change, the performance of the system remains rather stable thanks to the re-use of previously discovered visual routines while the visual exploration policy remains for long time sub-optimal. We conclude that the model has a robust behaviour but the acquisition of an optimal visual exploration policy is particularly hard given its complex dependence on statistical properties of the environment, showing another of the di?culties that adaptive active vision agents must face. (literal)

Prodotto di

• Tecnologie linguistiche e gestione della conoscenza (IC.P02.004.001) (Modulo)
• Institute of computational linguistics \"Antonio Zampolli\" (ILC) (Istituto)

Autore CNR

• GIOVANNI PEZZULO (Unità di personale interno)

Insieme di parole chiave

• Keywords of "Learning to Look in Different Environments: An Active-Vision Model which Learns and Readapts Visual Routines" (Insieme di parole chiave)

Incoming links:

Autore CNR di

• GIOVANNI PEZZULO (Unità di personale interno)

Prodotto

• Institute of computational linguistics \"Antonio Zampolli\" (ILC) (Istituto)
• Tecnologie linguistiche e gestione della conoscenza (IC.P02.004.001) (Modulo)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#rivistaDi

• Lecture notes in computer science (Rivista)

Insieme di parole chiave di

• Keywords of "Learning to Look in Different Environments: An Active-Vision Model which Learns and Readapts Visual Routines" (Insieme di parole chiave)