Development of Categorisation Abilities in Evolving Embodied Agents: A Study of Internal Representations with External Social Inputs

  • Francesco PuglieseEmail author


This paper investigates the behaviour of embodied and situated agents, which perform tasks requiring categorisation skills. These agents are simulated robots selected by an artificial adaptation process. Their task is to categorise objects with different shapes. To achieve this goal, the robots can use sensory information from the environment and external “linguistic” inputs. The results show that the agents are able to solve the categorisation task by conveniently integrating the experienced sensory-motor states and linguistic inputs. The aim of this work is to demonstrate that autonomous agents are able to develop some high-level cognitive abilities. Interestingly, the behavioural pattern seems to be in agreement with the theoretical hypothesis “social” information (external inputs) facilitates individual capacity to categorise, by producing good internal representations.


Target Area Hide Node Hide Neuron Social Information External Input 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.



I gratefully thank Stefano Nolfi, Davide Marocco, and Marco Mirolli for their helpful support.


  1. 1.
    Nolfi, S.: Categories formation in self-organizing embodied agents. In: Cohen, H., Lefebvre, C. (eds.) Handbook of Categorisation in Cognitive Science, pp. 869–889. Elsevier (2005)Google Scholar
  2. 2.
    Nolfi, S.: Behaviour as a complex adaptive system: on the role of self-organization in the development of individual and collective behaviour. ComplexUs 2(3–4), 195–203 (2006)Google Scholar
  3. 3.
    Marocco, D., Nolfi, S., Ikegami, T.: Emergence of protosentences in artificial communicating systems. IEEE Trans. Auton. Mental Dev. 3(2), 146–153 (2011)CrossRefGoogle Scholar
  4. 4.
    Tuci, E., Massera, G., Nolfi, S.: Active categorical perception in an evolved anthropomorphic robotic arm. In: IEEE Congress on Evolutionary Computation (CEC’09), pp. 31–38, IEEE, Trondheim (2009)Google Scholar
  5. 5.
    Kato, T., Floreano, D.: An evolutionary active-vision system. In: Proceedings of the Congress on Evolutionary Computation, vol. 1, pp. 107–114, IEEE, Seoul (2001)Google Scholar
  6. 6.
    Mirolli, M., Ferrauto, T., Nolfi, S.: Categorisation through evidence accumulation in an active vision system. Connect. Sci. 22(4), 331–354 (2010)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  1. 1.Natural and Artificial Cognition LaboratoryUniversity of NaplesNaplesItaly

Personalised recommendations