Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761026
Chie Hieida, Takato Horii, T. Nagai
In this paper, a model of emotions is proposed based on various neurological and psychological findings. The proposed model consists of three layers: the external/internal appraisal layer, the prediction/decision-making layer, and the emotional memory layer. We implement the proposed model by integrating some deep learning modules such as recurrent attention model, convolutional long short-term memory, and deep deterministic policy gradient. We set a “facial expression” task simulating mother-child interactions and verified emotion differentiation during the task. We also examine the trained model in the “still face” experiment. A claim in this study is that it is a very important step for the constructive approach to compare the proposed model with real human subjects in the same experiment that was carried out in the psychological studies.
{"title":"Toward Empathic Communication: Emotion Differentiation via Face-to-Face Interaction in Generative Model of Emotion","authors":"Chie Hieida, Takato Horii, T. Nagai","doi":"10.1109/DEVLRN.2018.8761026","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761026","url":null,"abstract":"In this paper, a model of emotions is proposed based on various neurological and psychological findings. The proposed model consists of three layers: the external/internal appraisal layer, the prediction/decision-making layer, and the emotional memory layer. We implement the proposed model by integrating some deep learning modules such as recurrent attention model, convolutional long short-term memory, and deep deterministic policy gradient. We set a “facial expression” task simulating mother-child interactions and verified emotion differentiation during the task. We also examine the trained model in the “still face” experiment. A claim in this study is that it is a very important step for the constructive approach to compare the proposed model with real human subjects in the same experiment that was carried out in the psychological studies.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114938727","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761015
Claus Lang, G. Schillaci, V. Hafner
This work investigates the role of predictive models in the implementation of basic cognitive skills in robots, such as the capability to distinguish between self-generated actions and those generated by other individuals and the capability to maintain an enhanced internal visual representation of the world, where objects covered by the robot's own body in the original image may be visible in the enhanced one. A developmental approach is adopted for this purpose. In particular, a humanoid robot is learning, through a self-exploration behaviour, the sensory consequences (in the visual domain) of self-generated movements. The generated sensorimotor experience is used as training data for a deep convolutional neural network that maps proprioceptive and motor data (e.g. initial arm joint positions and applied motor commands) onto the visual consequences of these actions. This forward model is then used in two experiments. First, for generating visual predictions of self-generated movements, which are compared to actual visual perceptions and then used to compute a prediction error. This error is shown to be higher when there is an external subject performing actions, compared to situations where the robot is observing only itself. This supports the idea that prediction errors may serve as a cue for distinguishing between self and other, a fundamental prerequisite for the sense of agency. Secondly, we show how predictions can be used to attenuate self-generated movements, and thus create enhanced visual perceptions, where the sight of objects - originally occluded by the robot body - is still maintained. This may represent an important tool both for cognitive development in robots and for the understanding of the sense of object permanence in humans.
{"title":"A Deep Convolutional Neural Network Model for Sense of Agency and Object Permanence in Robots","authors":"Claus Lang, G. Schillaci, V. Hafner","doi":"10.1109/DEVLRN.2018.8761015","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761015","url":null,"abstract":"This work investigates the role of predictive models in the implementation of basic cognitive skills in robots, such as the capability to distinguish between self-generated actions and those generated by other individuals and the capability to maintain an enhanced internal visual representation of the world, where objects covered by the robot's own body in the original image may be visible in the enhanced one. A developmental approach is adopted for this purpose. In particular, a humanoid robot is learning, through a self-exploration behaviour, the sensory consequences (in the visual domain) of self-generated movements. The generated sensorimotor experience is used as training data for a deep convolutional neural network that maps proprioceptive and motor data (e.g. initial arm joint positions and applied motor commands) onto the visual consequences of these actions. This forward model is then used in two experiments. First, for generating visual predictions of self-generated movements, which are compared to actual visual perceptions and then used to compute a prediction error. This error is shown to be higher when there is an external subject performing actions, compared to situations where the robot is observing only itself. This supports the idea that prediction errors may serve as a cue for distinguishing between self and other, a fundamental prerequisite for the sense of agency. Secondly, we show how predictions can be used to attenuate self-generated movements, and thus create enhanced visual perceptions, where the sight of objects - originally occluded by the robot body - is still maintained. This may represent an important tool both for cognitive development in robots and for the understanding of the sense of object permanence in humans.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129732790","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761016
J. Rączaszek-Leonardi, T. Deacon
The relation of symbolic cognition to embodied and situated bodily dynamics remains one of the hardest problems in the contemporary cognitive sciences. In this paper we show that one of the possible factors contributing to this difficulty is the way the problem is posed. Basing on the theoretical frameworks of cognitive semiotics, ecological psychology and dynamical systems we point to an alternative way of formulating the problem and show how it suggests possible novel solutions. We illustrate the usefulness of this theoretical change in the domain of language development and draw conclusions for computational models of the emergence of symbols in natural cognition and communication as well as in artificial systems.
{"title":"Ungrounding symbols in language development: implications for modeling emergent symbolic communication in artificial systems","authors":"J. Rączaszek-Leonardi, T. Deacon","doi":"10.1109/DEVLRN.2018.8761016","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761016","url":null,"abstract":"The relation of symbolic cognition to embodied and situated bodily dynamics remains one of the hardest problems in the contemporary cognitive sciences. In this paper we show that one of the possible factors contributing to this difficulty is the way the problem is posed. Basing on the theoretical frameworks of cognitive semiotics, ecological psychology and dynamical systems we point to an alternative way of formulating the problem and show how it suggests possible novel solutions. We illustrate the usefulness of this theoretical change in the domain of language development and draw conclusions for computational models of the emergence of symbols in natural cognition and communication as well as in artificial systems.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129902400","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761022
Rodrigo Zenha, Pedro Vicente, L. Jamone, A. Bernardino
The term ‘body schema’ refers to a computational representation of a physical body; the neural representation of a human body, or the numerical representation of a robot body. In both humans and robots, such a representation is crucial to accurately control body movements. While humans learn and continuously adapt their body schema based on multimodal perception and neural plasticity, robots are typically assigned with a fixed analytical model (e.g., the robot kinematics) which describes their bodies. However, there are always discrepancies between a model and the real robot, and they vary over time, thus affecting the accuracy of movement control. In this work, we equip a humanoid robot with the ability to incrementally estimate such model inaccuracies by touching known planar surfaces (e.g., walls) in its vicinity through motor babbling exploration, effectively adapting its own body schema based on the contact information alone. The problem is formulated as an adaptive parameter estimation (Extended Kalman Filter) which makes use of planar constraints obtained at each contact detection. We compare different incremental update methods through an extensive set of experiments with a realistic simulation of the iCub humanoid robot, showing that the model inaccuracies can be reduced by more than 80%.
{"title":"Incremental adaptation of a robot body schema based on touch events","authors":"Rodrigo Zenha, Pedro Vicente, L. Jamone, A. Bernardino","doi":"10.1109/DEVLRN.2018.8761022","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761022","url":null,"abstract":"The term ‘body schema’ refers to a computational representation of a physical body; the neural representation of a human body, or the numerical representation of a robot body. In both humans and robots, such a representation is crucial to accurately control body movements. While humans learn and continuously adapt their body schema based on multimodal perception and neural plasticity, robots are typically assigned with a fixed analytical model (e.g., the robot kinematics) which describes their bodies. However, there are always discrepancies between a model and the real robot, and they vary over time, thus affecting the accuracy of movement control. In this work, we equip a humanoid robot with the ability to incrementally estimate such model inaccuracies by touching known planar surfaces (e.g., walls) in its vicinity through motor babbling exploration, effectively adapting its own body schema based on the contact information alone. The problem is formulated as an adaptive parameter estimation (Extended Kalman Filter) which makes use of planar constraints obtained at each contact detection. We compare different incremental update methods through an extensive set of experiments with a realistic simulation of the iCub humanoid robot, showing that the model inaccuracies can be reduced by more than 80%.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126919254","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761021
Matthieu Zimmer, Y. Boniface, A. Dutech
In the framework of model-free deep reinforcement learning with continuous sensorimotor space, we propose a new type of transfer learning, inspired by the child development, where the sensorimotor space of an agent grows while it is learning a policy. To decide how the dimensions grow in our neural network based actor-critic, we add new developmental layers to the neural networks which progressively uncover some dimensions of the sensorimotor space following an Intrinsic Motivation heuristic. To mitigate the catastrophic forgetting problem, we take inspiration from the Elastic Weight Constraint to regulate the learning of the neural controller. We validate our approach using two state-of-the-art algorithms (DDPG and NFAC) on two high-dimensional environment benchmarks (Half-Cheetah and Humanoid). We show that searching first for a suboptimal solution in a subset of the parameter space, and then in the full space, is helpful to bootstrap learning algorithms, and thus reach better performances in fewer episodes.
{"title":"Developmental Reinforcement Learning through Sensorimotor Space Enlargement","authors":"Matthieu Zimmer, Y. Boniface, A. Dutech","doi":"10.1109/DEVLRN.2018.8761021","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761021","url":null,"abstract":"In the framework of model-free deep reinforcement learning with continuous sensorimotor space, we propose a new type of transfer learning, inspired by the child development, where the sensorimotor space of an agent grows while it is learning a policy. To decide how the dimensions grow in our neural network based actor-critic, we add new developmental layers to the neural networks which progressively uncover some dimensions of the sensorimotor space following an Intrinsic Motivation heuristic. To mitigate the catastrophic forgetting problem, we take inspiration from the Elastic Weight Constraint to regulate the learning of the neural controller. We validate our approach using two state-of-the-art algorithms (DDPG and NFAC) on two high-dimensional environment benchmarks (Half-Cheetah and Humanoid). We show that searching first for a suboptimal solution in a subset of the parameter space, and then in the full space, is helpful to bootstrap learning algorithms, and thus reach better performances in fewer episodes.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129076610","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761033
Qingpeng Zhu, Chong Zhang, J. Triesch, Bertram E. Shi
A central aspect of the development of visual perception is the autonomous calibration of various kinds of eye movements including saccadic, pursuit, or vergence eye movements. An important but less well-studied class of eye movements are so-called torsional eye movements, where the eyes rotate around the line of sight. In humans, such torsional eye movements obey certain lawful relationships such as Listing's Law. However, it is still an open question how these eye movements develop and what learning processes may contribute to their development. Here we propose a model of the development of torsional eye movements based on the active efficient coding (AEC) framework. AEC models the joint development of sensory encoding and movements of the sense organs to maximize the overall coding efficiency of the perceptual system. Our results demonstrate that optimizing coding efficiency in this way leads to torsional eye movements consistent with Listing's Law describing torsional eye movements in humans. This suggests that humanoid robots aiming to maximize the coding efficiency of their visual systems could also benefit from physical or simulated torsional eye movements.
{"title":"Autonomous learning of cyclovergence control based on Active Efficient Coding","authors":"Qingpeng Zhu, Chong Zhang, J. Triesch, Bertram E. Shi","doi":"10.1109/DEVLRN.2018.8761033","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761033","url":null,"abstract":"A central aspect of the development of visual perception is the autonomous calibration of various kinds of eye movements including saccadic, pursuit, or vergence eye movements. An important but less well-studied class of eye movements are so-called torsional eye movements, where the eyes rotate around the line of sight. In humans, such torsional eye movements obey certain lawful relationships such as Listing's Law. However, it is still an open question how these eye movements develop and what learning processes may contribute to their development. Here we propose a model of the development of torsional eye movements based on the active efficient coding (AEC) framework. AEC models the joint development of sensory encoding and movements of the sense organs to maximize the overall coding efficiency of the perceptual system. Our results demonstrate that optimizing coding efficiency in this way leads to torsional eye movements consistent with Listing's Law describing torsional eye movements in humans. This suggests that humanoid robots aiming to maximize the coding efficiency of their visual systems could also benefit from physical or simulated torsional eye movements.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132964540","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761027
V. Charisi, Cynthia C. S. Liem, E. Gómez
Humans have the capacity to invent novel ideas and to create new artifacts that affect the surrounding environment. However, it is unclear how this capacity emerges and develops in biological systems. This paper presents an empirical study which investigates the development of novelty-based cognitive processes in the context of unstructured music-making activities in early childhood. We used principles of intuitive theories of emergence, the paradigm of overlapping waves of mechanisms of change and theories of music cognitive development to theoretically conceptualize the developmental process in the specific context. We applied the methodological principles of micro-genetic analysis for the development of an annotation scheme of micro-behaviors, which correspond to a set of cognitive processes. We took into consideration child's behavioral manifestations of music-induced affective engagement, as an indicator of intrinsic motivation. Our results suggest that the process of transition from spontaneous towards deliberate actions develops through exploratory actions, evaluation of the outcomes, reasoning and planning. The structure of these actions appears in the form of dynamic overlapping waves rather than in a linear or iterative manner. Additionally, our results indicate that children in early years make use of the affordances of the provided tools to scaffold their transition from concrete visual representation of sonic features towards abstract musical thinking, which suggests that musical development appears with the generative tension between action and symbol. Implications and future work are discussed regarding the development of intelligent robotic systems for user adaptive scaffolding of the observed mechanisms of change.
{"title":"Novelty-based cognitive processes in unstructured music-making settings in early childhood","authors":"V. Charisi, Cynthia C. S. Liem, E. Gómez","doi":"10.1109/DEVLRN.2018.8761027","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761027","url":null,"abstract":"Humans have the capacity to invent novel ideas and to create new artifacts that affect the surrounding environment. However, it is unclear how this capacity emerges and develops in biological systems. This paper presents an empirical study which investigates the development of novelty-based cognitive processes in the context of unstructured music-making activities in early childhood. We used principles of intuitive theories of emergence, the paradigm of overlapping waves of mechanisms of change and theories of music cognitive development to theoretically conceptualize the developmental process in the specific context. We applied the methodological principles of micro-genetic analysis for the development of an annotation scheme of micro-behaviors, which correspond to a set of cognitive processes. We took into consideration child's behavioral manifestations of music-induced affective engagement, as an indicator of intrinsic motivation. Our results suggest that the process of transition from spontaneous towards deliberate actions develops through exploratory actions, evaluation of the outcomes, reasoning and planning. The structure of these actions appears in the form of dynamic overlapping waves rather than in a linear or iterative manner. Additionally, our results indicate that children in early years make use of the affordances of the provided tools to scaffold their transition from concrete visual representation of sonic features towards abstract musical thinking, which suggests that musical development appears with the generative tension between action and symbol. Implications and future work are discussed regarding the development of intelligent robotic systems for user adaptive scaffolding of the observed mechanisms of change.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115811782","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761036
Ryo Ozaki, T. Taniguchi
This paper describes an accelerated nonparametric Bayesian double articulation analyzer (NPB-DAA) for enabling a developmental robot to acquire words and phonemes directly from speech signals without labeled data in more realistic scenario than conventional NPB-DAA. Word discovery and phoneme acquisition are known as important tasks in human child development. Human infants can discover words and phonemes from raw speech signals at eight months without any label data, unlike supervised learning-based speech recognition systems. NPB-DAA was proposed by Taniguchi et al. and shown to be able to perform simultaneous word and phoneme discovery without any label data. However, the computational cost of NPB-DAA was extremely large, and thus could not be applied to large-scale speech data. In this paper, we introduce lookup tables for conventional NPB-DAA to reduce the computational cost and developed an accelerated NPB-DAA. Using the lookup tables, values calculated in each subroutine are memorized and reused in the subsequent calculations. This acceleration does not harm the quality of word and phoneme discovery because the introduction of the lookup tables is theoretically supported. This paper also shows that our accelerated NPB-DAA significantly reduced the computational cost by 90% compared to conventional NPB-DAA.
{"title":"Accelerated Nonparametric Bayesian Double Articulation Analyzer for Unsupervised Word Discovery","authors":"Ryo Ozaki, T. Taniguchi","doi":"10.1109/DEVLRN.2018.8761036","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761036","url":null,"abstract":"This paper describes an accelerated nonparametric Bayesian double articulation analyzer (NPB-DAA) for enabling a developmental robot to acquire words and phonemes directly from speech signals without labeled data in more realistic scenario than conventional NPB-DAA. Word discovery and phoneme acquisition are known as important tasks in human child development. Human infants can discover words and phonemes from raw speech signals at eight months without any label data, unlike supervised learning-based speech recognition systems. NPB-DAA was proposed by Taniguchi et al. and shown to be able to perform simultaneous word and phoneme discovery without any label data. However, the computational cost of NPB-DAA was extremely large, and thus could not be applied to large-scale speech data. In this paper, we introduce lookup tables for conventional NPB-DAA to reduce the computational cost and developed an accelerated NPB-DAA. Using the lookup tables, values calculated in each subroutine are memorized and reused in the subsequent calculations. This acceleration does not harm the quality of word and phoneme discovery because the introduction of the lookup tables is theoretically supported. This paper also shows that our accelerated NPB-DAA significantly reduced the computational cost by 90% compared to conventional NPB-DAA.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121305519","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761035
G. Baldassarre, Francesco Mannella, V. Santucci, E. Somogyi, Lisa Jacquey, Mollie Hamilton, J. O'Regan
Open-ended learning allows humans and robots to autonomously acquire an increasingly large repertoire of skills, that later can allow them to produce suitable actions to achieve desirable effects in the environment (‘goals'). Empirical evidence from developmental psychology suggests that a pivotal mechanism possibly driving open-ended learning is represented by action-outcome contingencies. Here we propose a specific hypothesis, expressed in the form of a blueprint cognitive architecture, that sketches the general mechanisms through which contingency-based open-ended learning might take place. According to this hypothesis, the matching (or distance) between a desired goal and the actual effect produced by the action can be used to drive the learning of both the motor skill used to accomplish the goal and the internal representation of the action outcome. We report here a computational model that implements the hypothesis and we illustrate two developmental psychology experiments related to the presented theory. Overall the model and experiments show the soundness of the hypothesis and represent a start towards validating it experimentally.
{"title":"Action-outcome contingencies as the engine of open-ended learning: computational models and developmental experiments","authors":"G. Baldassarre, Francesco Mannella, V. Santucci, E. Somogyi, Lisa Jacquey, Mollie Hamilton, J. O'Regan","doi":"10.1109/DEVLRN.2018.8761035","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761035","url":null,"abstract":"Open-ended learning allows humans and robots to autonomously acquire an increasingly large repertoire of skills, that later can allow them to produce suitable actions to achieve desirable effects in the environment (‘goals'). Empirical evidence from developmental psychology suggests that a pivotal mechanism possibly driving open-ended learning is represented by action-outcome contingencies. Here we propose a specific hypothesis, expressed in the form of a blueprint cognitive architecture, that sketches the general mechanisms through which contingency-based open-ended learning might take place. According to this hypothesis, the matching (or distance) between a desired goal and the actual effect produced by the action can be used to drive the learning of both the motor skill used to accomplish the goal and the internal representation of the action outcome. We report here a computational model that implements the hypothesis and we illustrate two developmental psychology experiments related to the presented theory. Overall the model and experiments show the soundness of the hypothesis and represent a start towards validating it experimentally.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"381 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122349769","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}
Pub Date : 2018-09-01DOI: 10.1109/DEVLRN.2018.8761018
D. Luo, Mengxi Nie, Tao Zhang, Xihong Wu
Robot reaching is a fundamental skill for knowing about the environment through interacting with objects and completing complex manipulation tasks. The topic has been studied widely for decades. In the paper, with reference to the relevant mechanism of human, a novel strategy for developing robot reaching skill is proposed, in which the whole process is divided into two stages including rough reaching and iterative adjustment. Generally in the process of obtaining spatial information of target object, the accuracy of the absolute positioning might be severely affected due to inevitable errors derived from sensing means (e.g. camera) in real world scenario. On the contrary, the accuracy of relative positioning will be much better, in which we only require answering the relative location between the target and the end-effector. Under this view, the proposed method, called the relative-location based approximating strategy (RLA), firstly attempts to move the end-effector to the target roughly with a simple inverse model, and then gradually approximates to the target according to the information of the relative location, i.e. the direction of the target relative to the end-effector. To accomplish such an approximating process, an internal model regarding to base directions is developed, where the motor babbling is involved under the inspiration of infants development mechanism. The approach was experimentally validated using the child-sized physical humanoid robot PKU-HR6.0II in a completely autonomous style and the results illustrate the effectiveness and superiority of the proposed strategy.
{"title":"Developing Robot Reaching Skill with Relative-Location based Approximating","authors":"D. Luo, Mengxi Nie, Tao Zhang, Xihong Wu","doi":"10.1109/DEVLRN.2018.8761018","DOIUrl":"https://doi.org/10.1109/DEVLRN.2018.8761018","url":null,"abstract":"Robot reaching is a fundamental skill for knowing about the environment through interacting with objects and completing complex manipulation tasks. The topic has been studied widely for decades. In the paper, with reference to the relevant mechanism of human, a novel strategy for developing robot reaching skill is proposed, in which the whole process is divided into two stages including rough reaching and iterative adjustment. Generally in the process of obtaining spatial information of target object, the accuracy of the absolute positioning might be severely affected due to inevitable errors derived from sensing means (e.g. camera) in real world scenario. On the contrary, the accuracy of relative positioning will be much better, in which we only require answering the relative location between the target and the end-effector. Under this view, the proposed method, called the relative-location based approximating strategy (RLA), firstly attempts to move the end-effector to the target roughly with a simple inverse model, and then gradually approximates to the target according to the information of the relative location, i.e. the direction of the target relative to the end-effector. To accomplish such an approximating process, an internal model regarding to base directions is developed, where the motor babbling is involved under the inspiration of infants development mechanism. The approach was experimentally validated using the child-sized physical humanoid robot PKU-HR6.0II in a completely autonomous style and the results illustrate the effectiveness and superiority of the proposed strategy.","PeriodicalId":236346,"journal":{"name":"2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)","volume":"320 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114205037","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：