Chained learning architectures in a simple closed-loop behavioural context

Tomas Kulvicius, Bernd Porr, Florentin Wörgötter

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

Resumé

OBJECTIVE: Living creatures can learn or improve their behaviour by temporally correlating sensor cues where near-senses (e.g., touch, taste) follow after far-senses (vision, smell). Such type of learning is related to classical and/or operant conditioning. Algorithmically all these approaches are very simple and consist of single learning unit. The current study is trying to solve this problem focusing on chained learning architectures in a simple closed-loop behavioural context.

METHODS: We applied temporal sequence learning (Porr B and Wörgötter F 2006) in a closed-loop behavioural system where a driving robot learns to follow a line. Here for the first time we introduced two types of chained learning architectures named linear chain and honeycomb chain. We analyzed such architectures in an open and closed-loop context and compared them to the simple learning unit.

CONCLUSIONS: By implementing two types of simple chained learning architectures we have demonstrated that stable behaviour can also be obtained in such architectures. Results also suggest that chained architectures can be employed and better behavioural performance can be obtained compared to simple architectures in cases where we have sparse inputs in time and learning normally fails because of weak correlations.

OriginalsprogEngelsk
TidsskriftBiological Cybernetics
Vol/bind97
Udgave nummer5-6
Sider (fra-til)363-78
Antal sider16
ISSN0340-1200
DOI
StatusUdgivet - dec. 2007

Fingeraftryk

Robots
Sensors
Operant Conditioning
Classical Conditioning
Smell
Cues

Citer dette

Kulvicius, Tomas ; Porr, Bernd ; Wörgötter, Florentin. / Chained learning architectures in a simple closed-loop behavioural context. I: Biological Cybernetics. 2007 ; Bind 97, Nr. 5-6. s. 363-78.
@article{552af33e5eb64b74a639900f3aeded8c,
title = "Chained learning architectures in a simple closed-loop behavioural context",
abstract = "OBJECTIVE: Living creatures can learn or improve their behaviour by temporally correlating sensor cues where near-senses (e.g., touch, taste) follow after far-senses (vision, smell). Such type of learning is related to classical and/or operant conditioning. Algorithmically all these approaches are very simple and consist of single learning unit. The current study is trying to solve this problem focusing on chained learning architectures in a simple closed-loop behavioural context.METHODS: We applied temporal sequence learning (Porr B and W{\"o}rg{\"o}tter F 2006) in a closed-loop behavioural system where a driving robot learns to follow a line. Here for the first time we introduced two types of chained learning architectures named linear chain and honeycomb chain. We analyzed such architectures in an open and closed-loop context and compared them to the simple learning unit.CONCLUSIONS: By implementing two types of simple chained learning architectures we have demonstrated that stable behaviour can also be obtained in such architectures. Results also suggest that chained architectures can be employed and better behavioural performance can be obtained compared to simple architectures in cases where we have sparse inputs in time and learning normally fails because of weak correlations.",
keywords = "Animals, Artificial Intelligence, Behavior, Computer Simulation, Humans, Learning, Models, Biological, Neural Networks (Computer), Pattern Recognition, Automated, Robotics",
author = "Tomas Kulvicius and Bernd Porr and Florentin W{\"o}rg{\"o}tter",
year = "2007",
month = "12",
doi = "10.1007/s00422-007-0176-y",
language = "English",
volume = "97",
pages = "363--78",
journal = "Biological Cybernetics",
issn = "0340-1200",
publisher = "Heinemann",
number = "5-6",

}

Chained learning architectures in a simple closed-loop behavioural context. / Kulvicius, Tomas; Porr, Bernd; Wörgötter, Florentin.

I: Biological Cybernetics, Bind 97, Nr. 5-6, 12.2007, s. 363-78.

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

TY - JOUR

T1 - Chained learning architectures in a simple closed-loop behavioural context

AU - Kulvicius, Tomas

AU - Porr, Bernd

AU - Wörgötter, Florentin

PY - 2007/12

Y1 - 2007/12

N2 - OBJECTIVE: Living creatures can learn or improve their behaviour by temporally correlating sensor cues where near-senses (e.g., touch, taste) follow after far-senses (vision, smell). Such type of learning is related to classical and/or operant conditioning. Algorithmically all these approaches are very simple and consist of single learning unit. The current study is trying to solve this problem focusing on chained learning architectures in a simple closed-loop behavioural context.METHODS: We applied temporal sequence learning (Porr B and Wörgötter F 2006) in a closed-loop behavioural system where a driving robot learns to follow a line. Here for the first time we introduced two types of chained learning architectures named linear chain and honeycomb chain. We analyzed such architectures in an open and closed-loop context and compared them to the simple learning unit.CONCLUSIONS: By implementing two types of simple chained learning architectures we have demonstrated that stable behaviour can also be obtained in such architectures. Results also suggest that chained architectures can be employed and better behavioural performance can be obtained compared to simple architectures in cases where we have sparse inputs in time and learning normally fails because of weak correlations.

AB - OBJECTIVE: Living creatures can learn or improve their behaviour by temporally correlating sensor cues where near-senses (e.g., touch, taste) follow after far-senses (vision, smell). Such type of learning is related to classical and/or operant conditioning. Algorithmically all these approaches are very simple and consist of single learning unit. The current study is trying to solve this problem focusing on chained learning architectures in a simple closed-loop behavioural context.METHODS: We applied temporal sequence learning (Porr B and Wörgötter F 2006) in a closed-loop behavioural system where a driving robot learns to follow a line. Here for the first time we introduced two types of chained learning architectures named linear chain and honeycomb chain. We analyzed such architectures in an open and closed-loop context and compared them to the simple learning unit.CONCLUSIONS: By implementing two types of simple chained learning architectures we have demonstrated that stable behaviour can also be obtained in such architectures. Results also suggest that chained architectures can be employed and better behavioural performance can be obtained compared to simple architectures in cases where we have sparse inputs in time and learning normally fails because of weak correlations.

KW - Animals

KW - Artificial Intelligence

KW - Behavior

KW - Computer Simulation

KW - Humans

KW - Learning

KW - Models, Biological

KW - Neural Networks (Computer)

KW - Pattern Recognition, Automated

KW - Robotics

U2 - 10.1007/s00422-007-0176-y

DO - 10.1007/s00422-007-0176-y

M3 - Journal article

C2 - 17912544

VL - 97

SP - 363

EP - 378

JO - Biological Cybernetics

JF - Biological Cybernetics

SN - 0340-1200

IS - 5-6

ER -