Repository | Book | Chapter
(2015) Artificial life and computational intelligence, Dordrecht, Springer.
Exploring the periphery of knowledge by intrinsically motivated systems
Kirill Makukhin , Scott Bolland
pp. 49-61
Intrinsically motivated learning is essential for the development of a wide range of competences. However, the neural substrate for the motivational signal as well as how this signal facilitates the processes of building competences are poorly understood. In this paper we exploit a biologically plausible approach, showing that an intrinsically motivated system where the motivation depends on stimulus familiarity as an inverted U-shape, exhibits well-structured exploration behaviour. Furthermore, we show that such behaviour may lead to the emergence of complex competences such as object affordances.
Publication details
DOI: 10.1007/978-3-319-14803-8_4
Full citation:
Makukhin, K. , Bolland, S. (2015)., Exploring the periphery of knowledge by intrinsically motivated systems, in M. Randall (ed.), Artificial life and computational intelligence, Dordrecht, Springer, pp. 49-61.