TY - GEN
T1 - Are there representations in embodied evolved agents? Taking measures
AU - Avraham, Hezi
AU - Chechik, Gal
AU - Ruppin, Eytan
N1 - Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2003.
PY - 2003
Y1 - 2003
N2 - The question of conceptual representation has received considerable attention in philosophy, neuroscience and embodied evolved agents. Numerous theories on the interpretation of the term 'representation' exist, and many arguments have been made for and against the existence of representations in animate and animat agents. Our work studies this question in evolved artificial embodied agents in a quantitatively rigorous manner, for the first time. We develop two measures, based on information theory, to account for representations. These measures are studied by applying them to evolved agents performing a visual categorization, generalized XOR task. Our results show that having quantitative measures still leaves one with arbitrary "threshold values" decisions which permit wide freedom in determining the existence of representations. However, and more importantly, our results show that information-theoretic measures can still be used efficiently to identify discriminative neural patterns and internal structures that characterize a representation, if the latter is formed.
AB - The question of conceptual representation has received considerable attention in philosophy, neuroscience and embodied evolved agents. Numerous theories on the interpretation of the term 'representation' exist, and many arguments have been made for and against the existence of representations in animate and animat agents. Our work studies this question in evolved artificial embodied agents in a quantitatively rigorous manner, for the first time. We develop two measures, based on information theory, to account for representations. These measures are studied by applying them to evolved agents performing a visual categorization, generalized XOR task. Our results show that having quantitative measures still leaves one with arbitrary "threshold values" decisions which permit wide freedom in determining the existence of representations. However, and more importantly, our results show that information-theoretic measures can still be used efficiently to identify discriminative neural patterns and internal structures that characterize a representation, if the latter is formed.
UR - http://www.scopus.com/inward/record.url?scp=7444245197&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-39432-7_80
DO - 10.1007/978-3-540-39432-7_80
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:7444245197
SN - 3540200576
SN - 9783540200574
T3 - Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)
SP - 743
EP - 752
BT - Advances in Artificial Life
A2 - Banzhaf, Wolfgang
A2 - Ziegler, Jens
A2 - Christaller, Thomas
A2 - Dittrich, Peter
A2 - Kim, Jan T.
PB - Springer Verlag
T2 - 7th European Conference on Artificial Life, ECAL 2003
Y2 - 14 September 2003 through 17 September 2003
ER -