While today's robots are able to perform sophisticated tasks, they can only act on objects they have been trained to recognize. This is a severe limitation: any robot will inevitably see new objects in unconstrained settings, and thus will always have visual knowledge gaps. However, standard visual modules are usually built on a limited set of classes and are based on the strong prior that an object must belong to one of those classes. Identifying whether an instance does not belong to the set of known categories (i.e. open set recognition), only partially tackles this problem, as a truly autonomous agent should be able not only to detect what it does not know, but also to extend dynamically its knowledge about the world. We contribute to this challenge with a deep learning architecture that can dynamically update its known classes in an end-to-end fashion. The proposed deep network, based on a deep extension of a non-parametric model, detects whether a perceived object belongs to the set of categories known by the system and learns it without the need to retrain the whole system from scratch. Annotated images about the new category can be provided by an `oracle' (i.e. human supervision), or by autonomous mining of the Web. Experiments on two different databases and on a robot platform demonstrate the promise of our approach.

Knowledge is Never Enough: Towards Web Aided Deep Open World Recognition / Mancini, Massimiliano; Karaoguz, Hakan; Ricci, Elisa; Jensfelt, Patric; Caputo, Barbara. - (2019), pp. 9537-9543. (Intervento presentato al convegno 2019 International Conference on Robotics and Automation (ICRA) tenutosi a Montreal, QC; Canada) [10.1109/ICRA.2019.8793803].

Knowledge is Never Enough: Towards Web Aided Deep Open World Recognition

Massimiliano Mancini
;
Barbara Caputo
2019

Abstract

While today's robots are able to perform sophisticated tasks, they can only act on objects they have been trained to recognize. This is a severe limitation: any robot will inevitably see new objects in unconstrained settings, and thus will always have visual knowledge gaps. However, standard visual modules are usually built on a limited set of classes and are based on the strong prior that an object must belong to one of those classes. Identifying whether an instance does not belong to the set of known categories (i.e. open set recognition), only partially tackles this problem, as a truly autonomous agent should be able not only to detect what it does not know, but also to extend dynamically its knowledge about the world. We contribute to this challenge with a deep learning architecture that can dynamically update its known classes in an end-to-end fashion. The proposed deep network, based on a deep extension of a non-parametric model, detects whether a perceived object belongs to the set of categories known by the system and learns it without the need to retrain the whole system from scratch. Annotated images about the new category can be provided by an `oracle' (i.e. human supervision), or by autonomous mining of the Web. Experiments on two different databases and on a robot platform demonstrate the promise of our approach.
2019
2019 International Conference on Robotics and Automation (ICRA)
data mining;learning (artificial intelligence);mobile robots;object recognition;robot vision;deep learning architecture;deep network;deep extension;nonparametric model;autonomous mining;robot platform;deep open world recognition;visual knowledge gaps;open set recognition;visual modules;Robots;Visualization;Training;Feature extraction;Task analysis;Artificial neural networks;Semantics
04 Pubblicazione in atti di convegno::04b Atto di convegno in volume
Knowledge is Never Enough: Towards Web Aided Deep Open World Recognition / Mancini, Massimiliano; Karaoguz, Hakan; Ricci, Elisa; Jensfelt, Patric; Caputo, Barbara. - (2019), pp. 9537-9543. (Intervento presentato al convegno 2019 International Conference on Robotics and Automation (ICRA) tenutosi a Montreal, QC; Canada) [10.1109/ICRA.2019.8793803].
File allegati a questo prodotto
File Dimensione Formato  
Mancini_Knowledge-is-Never_2019.pdf

solo gestori archivio

Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 320.11 kB
Formato Adobe PDF
320.11 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/11573/1331861
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 18
  • ???jsp.display-item.citation.isi??? 18
social impact