2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)最新文献

Development of compositional and contextual communication of robots by using the multiple timescales dynamic neural network 基于多时间尺度动态神经网络的机器人组成和上下文通信研究

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346137

Gibeom Park, J. Tani

{"title":"Development of compositional and contextual communication of robots by using the multiple timescales dynamic neural network","authors":"Gibeom Park, J. Tani","doi":"10.1109/DEVLRN.2015.7346137","DOIUrl":"https://doi.org/10.1109/DEVLRN.2015.7346137","url":null,"abstract":"The current paper introduces neurorobotics experiment on acquisition of complex communicative skills with human via learning. A dynamic neural network model which is characterized by its multiple timescale dynamics characteristics was utilized as a neuronal model for controlling a humanoid robot. In the experimental task, the humanoid robot was trained to generate specific sequential movement patterns as responding to various sequences of imperative gesture patterns demonstrated by the human subjects by following predefined compositional semantic rules. The experimental results showed that (1) the MTRNN can learn to extract compositional semantic rules with generalization in the higher cognitive level, (2) the MTRNN can develop further higher-order cognition capability for controlling the internal contextual processes as situated to on-going task sequences without being provided with cues for explicitly indicating task segmentation points. The analysis on the dynamic characteristics developed in the MTRNN through learning indicated that the aforementioned cognitive mechanisms were achieved by developing adequate functional hierarchy by utilizing the constraint of the multiple timescale property and the topological connectivity imposed on the network configuration.","PeriodicalId":164756,"journal":{"name":"2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128485470","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Active learning strategies and active control of complexity growth in naming games 命名游戏中主动学习策略与复杂性增长的主动控制

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346144

William Schueller, Pierre-Yves Oudeyer

引用次数: 11

Biologically inspired incremental learning for high-dimensional spaces 生物学启发的高维空间增量学习

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346155

A. Gepperth, Thomas Hecht, Mathieu Lefort, Ursula Körner

{"title":"Biologically inspired incremental learning for high-dimensional spaces","authors":"A. Gepperth, Thomas Hecht, Mathieu Lefort, Ursula Körner","doi":"10.1109/DEVLRN.2015.7346155","DOIUrl":"https://doi.org/10.1109/DEVLRN.2015.7346155","url":null,"abstract":"We propose an incremental, highly parallelizable, and constant-time complexity neural learning architecture for multi-class classification (and regression) problems that remains resource-efficient even when the number of input dimensions is very high (≥ 1000). This so-called projection-prediction (PRO-PRE) architecture is strongly inspired by biological information processing in that it uses a prototype-based, topologically organized hidden layer that updates hidden layer weights whenever an error occurs. The employed self-organizing map (SOM) learning adapts only the weights of localized neural sub-populations that are similar to the input, which explicitly avoids the catastrophic forgetting effect of MLPs in case new input statistics are presented. The readout layer applies linear regression to hidden layer activities subjected to a transfer function, making the whole system capable of representing strongly non-linear decision boundaries. The resource-efficiency of the algorithm stems from approximating similarity in the input space by proximity in the SOM layer due to the topological SOM projection property. This avoids the storage of inter-cluster distances (quadratic in number of hidden layer elements) or input space covariance matrices (quadratic in input dimensions) as other incremental algorithms typically do. Tests on the popular MNIST handwritten digit benchmark show that the algorithm compares favorably to state-of-the-art results, and parallelizability is demonstrated by analyzing the efficiency of a parallel GPU implementation of the architecture.","PeriodicalId":164756,"journal":{"name":"2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"101 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122816628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 9

Diversity-driven selection of exploration strategies in multi-armed bandits 多武装盗匪的多样性驱动勘探策略选择

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/devlrn.2015.7346130

Fabien C. Y. Benureau, Pierre-Yves Oudeyer

引用次数: 7

Incremental grounded language learning in robot-robot interactions — Examples from spatial language 机器人-机器人互动中的增量基础语言学习-来自空间语言的例子

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346140

Michael Spranger

引用次数: 13

To hear and to hold: Maternal naming and infant object exploration 听和抱:母亲的命名和婴儿的对象探索

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346125

Lucas Chang, K. D. Barbaro, G. Deák

引用次数: 1

Babybot challenge: Motor skills 婴儿挑战:运动技能

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346114

Patricia Shaw, Daniel Lewkowicz, Alexandros Giagkos, J. Law, Suresh Kumar, Mark H. Lee, Q. Shen

引用次数: 6

Using spatial representations in gesture to facilitate early word learning: A priming process model 使用手势中的空间表征促进早期单词学习:启动过程模型

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346117

J. Trafton, Anthony M. Harrison, W. Lawson

引用次数: 0

Towards hierarchical curiosity-driven exploration of sensorimotor models 对层次好奇驱动的感觉运动模型探索

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346146

Sébastien Forestier, Pierre-Yves Oudeyer

{"title":"Towards hierarchical curiosity-driven exploration of sensorimotor models","authors":"Sébastien Forestier, Pierre-Yves Oudeyer","doi":"10.1109/DEVLRN.2015.7346146","DOIUrl":"https://doi.org/10.1109/DEVLRN.2015.7346146","url":null,"abstract":"Curiosity-driven exploration mechanisms have been proposed to allow robots to actively explore high dimensional sensorimotor spaces in an open-ended manner [1], [2]. In such setups, competence-based intrinsic motivations show better results than knowledge-based exploration mechanisms which only monitor the learner's prediction performance [2], [3]. With competence-based intrinsic motivations, the learner explores its sensor space with a bias toward regions which are predicted to yield a high competence progress. Also, throughout its life, a developmental robot has to incrementally explore skills that add up to the hierarchy of previously learned skills, with a constraint being the cost of experimentation. Thus, a hierarchical exploration architecture could allow to reuse the sensorimotor models previously explored and to combine them to explore more efficiently new complex sensorimotor models. Here, we rely more specifically on the R-IAC and SAGG-RIAC series of architectures [3]. These architectures allow the learning of a single mapping between a motor and a sensor space with a competence-based intrinsic motivation. We describe some ways to extend these architectures with different tasks spaces that can be explored in a hierarchical manner, and mechanisms to handle this hierarchy of sensorimotor models that all need to be explored with an adequate amount of trials. We also describe an interactive task to evaluate the hierarchical learning mechanisms, where a robot has to explore its motor space in order to push an object to different locations. The robot can first explore how to make movements with its hand and then reuse this skill to explore the task of pushing an object.","PeriodicalId":164756,"journal":{"name":"2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131892016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 6

Seeing [u] aids vocal learning: Babbling and imitation of vowels using a 3D vocal tract model, reinforcement learning, and reservoir computing Seeing [u]帮助语音学习:使用3D声道模型、强化学习和存储库计算来咿呀学语和模仿元音

2015 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) Pub Date : 2015-12-07 DOI: 10.1109/DEVLRN.2015.7346142

M. Murakami, B. Kröger, P. Birkholz, J. Triesch

引用次数: 26