Etienne Dumesnil, Philippe-Olivier Beaulieu, M. Boukadoum
{"title":"Robotic implementation of classical and operant conditioning within a single SNN architecture","authors":"Etienne Dumesnil, Philippe-Olivier Beaulieu, M. Boukadoum","doi":"10.1109/ICCI-CC.2016.7862055","DOIUrl":null,"url":null,"abstract":"This work presents the implementation of operant conditioning (OC) and classical conditioning (CC) with a single spiking neural network (SNN) architecture, thus suggesting that the two types of leaning may relate to the same cognitive process. Both are achieved by using a modified version of spike-timing-dependent plasticity (STDP), where the connection weight between a cue neuron and an action neuron depends on the temporal relation between their spikes and those of a reward neuron. This reward driven STDP (RD-STDP) was implemented with simple computational resources to form an electronic robot's brain, using an adaptation of the synapto-dendritic kernel adapting neuron (SKAN) model. Then, a robot driven by the new neuronal architecture was tested in a maze with changing features, successfully exhibiting CC and OC. These results and the simple computational resources used make the proposed architecture promising for very large scale time-dependent parallel data analysis, with high capacity of adaptation in a dynamic environment. Moreover, it proposes a theoretic framework to model learning by conditioning.","PeriodicalId":135701,"journal":{"name":"2016 IEEE 15th International Conference on Cognitive Informatics & Cognitive Computing (ICCI*CC)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 IEEE 15th International Conference on Cognitive Informatics & Cognitive Computing (ICCI*CC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICCI-CC.2016.7862055","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 3
Abstract
This work presents the implementation of operant conditioning (OC) and classical conditioning (CC) with a single spiking neural network (SNN) architecture, thus suggesting that the two types of leaning may relate to the same cognitive process. Both are achieved by using a modified version of spike-timing-dependent plasticity (STDP), where the connection weight between a cue neuron and an action neuron depends on the temporal relation between their spikes and those of a reward neuron. This reward driven STDP (RD-STDP) was implemented with simple computational resources to form an electronic robot's brain, using an adaptation of the synapto-dendritic kernel adapting neuron (SKAN) model. Then, a robot driven by the new neuronal architecture was tested in a maze with changing features, successfully exhibiting CC and OC. These results and the simple computational resources used make the proposed architecture promising for very large scale time-dependent parallel data analysis, with high capacity of adaptation in a dynamic environment. Moreover, it proposes a theoretic framework to model learning by conditioning.