检索结果-内蒙古大学图书馆

Workshop on Electrical Engineering and Automation

作者： Hu, Lin Chao, Fei Jiang, Min Shi, MingHui Wang, Pan Xiamen Univ Dept Cognit Sci Fujian Key Lab Brain Like Intelligent Syst Xiamen Fujian Peoples R China

ISBN: (纸本)9783642279508

Inspired by infant development and brain research, this paper, combining the two research hotspot: developmental robotics and robotic hand-eye coordination, proposes to employ incremental process to implement robotic hand-eye coordination using double-neural networks under the experimental condition of two cameras and a robotic arm. This setup is to imitate the developmental progress of human infant to capture objects: and the networks are implemented by the constructive network which can simulate infant's brain development. Furthermore, this paper delicately describes the details of the experiment plan and steps to show the combining course of the two hotspot.

关键词： developmental robotics Robotic Hand-eye Coordination Constructive Network

来源：评论

学校读者我要写书评

暂无评论

Learning while Sleeping: Integrating Sleep-Inspired Consolidation with Human Feedback Learning

Learning while Sleeping: Integrating Sleep-Inspired Consolid...

引用

IEEE International Conference on Development and Learning (ICDL)

作者： Tarakli, Imene Di Nuovo, Alessandro Sheffield Hallam Univ Sheffield S Yorkshire England

ISBN: (纸本)9798350348569;9798350348552

Sleep plays a vital role in developmental learning. It allows the brain to consolidate daily learning experiences by replaying the memories accumulated throughout the day. In this work, we take inspiration from sleep and propose the Inverse Forward Offline Reinforcement Model (INFORM), a novel scalable framework that first learns a set of behaviours from human evaluative feedback, then consolidates the learning by applying an offline inverse reinforcement learning to the memorised trajectories. Experimental results demonstrate that INFORM is a feedback-efficient method that effectively learns an optimal policy that aligns with the intended behaviour of the human. A comparative analysis shows that the learnt policies are robust to dynamic changes in the environment and the recovered rewards allow the robot to be autonomous in its learning. Project website: https://***/view/inform-framework

关键词： developmental robotics Cognitive robotics Interactive Agents

来源：评论

学校读者我要写书评

暂无评论

Motor skill acquisition under environmental perturbations: On the necessity of alternate freezing and freeing of degrees of freedom

引用

ADAPTIVE BEHAVIOR 2004年第1期12卷 47-64页

作者： Berthouze, L Lungarella, M Neurosci Res Inst AIST Tsukuba Ibaraki 3058568 Japan

In a recent study on the pendulation of a small-sized humanoid robot (Lungarella & Berthouze, 2002a,b), we provided experimental evidence that starting with fewer degrees of freedom enables a more efficient exploration of the sensorimotor space during the acquisition of a task. The study came as support for the well-established framework of Bernstein (1967), namely that of an initial freezing of the distal degrees of freedom, followed by their progressive release and the exploitation of environmental and body dynamics. In this paper, we revisit our study by introducing a nonlinear coupling between environment and system. Under otherwise unchanged experimental conditions, we show that a single phase of freezing and subsequent freeing of degrees of freedom is not sufficient to achieve optimal performance, and instead, alternate freezing and freeing of degrees of freedom is required. The interest of this result is twofold: (1) it confirms the recent observation by Newell & Vaillancourt (2001) that Bernstein's (1967) framework may be too narrow to account for real data;(2) it suggests that perturbations that push the system outside its postural stability or increase the task complexity may be the mechanism that triggers alternate freezing and freeing of degrees of freedom.

关键词： developmental robotics embodiment neural oscillator freezing and freeing of degrees of freedom entrainment

来源：评论

学校读者我要写书评

暂无评论

Towards Learning the Foundations of Manipulation Actions with Unguided Exploration

Towards Learning the Foundations of Manipulation Actions wit...

引用

作者： Juett, Jonathan Emerson University of Michigan

学位级别：Ph.D., Doctor of Philosophy

Human infants are not born with the ability to reach and grasp. But after months of typical development, infants are capable of reaching and grasping reliably. During this time, the infant receives minimal guidance and learns primarily by observing its autonomous experience with its developing senses. How is it possible for this learning phenomenon to occur, especially when this experience begins with seemingly random motions? We present a computational model that allows an embodied robotic agent to learn these foundational actions in a manner consistent with infant learning. By examining the model and the resulting behaviors, we can identify knowledge sufficient to perform these actions, and how this knowledge may be represented. Our agent uses a graph representation for peripersonal space, the space surrounding the agent and in reach of its manipulators. The agent constructs the Peripersonal Space (PPS) Graph by performing random motions. These motions are performed with the table but no other nonself foreground objects present to facilitate simple image segmentation of an unoccluded view of the hand. For each pose visited, a node stores the joint angles that produced it and an image of the arm in this configuration. Edges connect each pair of nodes that have a feasible motion between them. Later in the learning process, the agent may used learned criteria to temporarily remove a node or edge from consideration if motion to it or along it is expected to cause a collision given the current position of a foreground object being treated as an obstacle. The PPS Graph provides a mapping between configuration space and image space, and the agent learns to use it as a powerful tool for planning manipulation actions. Initially, the only known actions are moves to selected PPS Graph nodes. The agent begins learning by executing move actions, and will continue to learn by applying the same learning method with other actions once it has defined them. The agent selects a rand

关键词： Sensorimotor learning Autonomous robot learning Peripersonal space Reaching and grasping Intrinsic motivation developmental robotics

来源：评论

学校读者我要写书评

暂无评论

Interactive learning of words and objects for a humanoid robot

Interactive learning of words and objects for a humanoid rob...

引用

作者： Yuxin Chen ENSTA-ParisTech

学位级别：博士

Future applications of robotics will re- quire adaptability to the environment, particularly the ability to recognize new objects through interac- tion with humans. In fact, two year old children have an impressive ability to learn to recognize new ob- jects via interaction with adults and without precise supervision. Therefore, following the developmental robotics approach, we develop in the thesis learning approaches for objects, associating their names and corresponding features, inspired by the capabilities of infants that are used during ambiguous interac- tion with parents. Under the framework of cross-situational learning, two latent topic discovery approaches, that is Non Negative Matrix Factorization (NMF) and Latent Dirichlet Association (LDA), are utilized to imple- ment multi-modal concept discovery, finding the un- derlying regularities in the raw dataflow to produce sets of words and their associated visual meanings, eg. the name of an object and its shape, or a color adjective and its correspondence in images, despite referential ambiguities and linguistic ambiguities. Besides, two active learning strategies: Maximum Reconstruction Error Based Selection (MRES) and Confidence Based Exploration (CBE), are proposed to improve incremental learning in terms of quality and speed. Finally, learning behaviors are compa- red between humans and the proposed models not only through extensive simulations but also through applications in real human-robot interactions.

关键词： developmental robotics word-referent learning cross-situational learning active learning Non negative Matrix Factorization (NMF) Latent Dirichlet Association (LDA)

来源：评论

学校读者我要写书评

暂无评论

Fast Online Model Learning for Controlling Complex Real-World Robots

Fast Online Model Learning for Controlling Complex Real-Worl...

引用

作者： Loviken, Pontus University of Plymouth

学位级别：博士

How can real robots with many degrees of freedom - without previous knowledge of them- selves or their environment - act and use the resulting observations to efficiently develop the ability to generate a wide set of useful behaviours? This thesis presents a novel framework that enables physical robots with many degrees of freedom to rapidly learn models for control from scratch. This can be done in previously inaccessible problem domains characterised by a lack of direct mappings from motor actions to outcomes, as well as state and action spaces too large for the full forward dynamics to be learned and used explicitly. The proposed framework is able to cope with these issues by the use of a set of local Goal Babbling models, that maps every outcome in a low dimensional task space to a specific action, together with a sparse higher level Reinforcement Learning model, that learns to navigate between the contexts from which each Goal Babbling model can be used. The two types of models can then be learned online an in parallel, using only the data a robot can collect by interacting with its environment. To show the potential of the approach we present two possible implementations of the frame- work, over two separate robot platforms: a simulated planar arm with up to 1, 000 degrees of freedom, and a real humanoid robot with 25 degrees of freedom. The results show that learning is rapid and essentially unaffected by the number of degrees of freedom of the robot, allowing for the generation of complex behaviours and skills after a relatively short training time. The planar arm is able to strategically plan series of motions in order to move its end-effector between any two parts of a crowded environment, within 10, 000 iterations. The humanoid robot is able to freely transition between states such as lying on the back, belly, and sides, and occasionally also sitting up, within only 1, 000 iterations. This corresponds to 30 − 60 minutes of real-world interactions. The mai

关键词： model learning Reinforcement learning Online learning Goal babbling inverse models Micro data learning developmental robotics real-world robots sensorimotor control

来源：评论

学校读者我要写书评

暂无评论

Generalising Predictable Object Movements Through Experience Using Schemas 14th

Generalising Predictable Object Movements Through Experience...

引用

14th International Conference on the Simulation of Adaptive Behavior (SAB)

作者： Kumar, Suresh Shaw, Patricia Lewkowicz, Daniel Giagkos, Alexandros Lee, Mark Shen, Qiang Aberystwyth Univ Dept Comp Sci Aberystwyth Dyfed Wales Sukkur Inst Business Adm Sukkur IBA Sukkur Pakistan

ISBN: (纸本)9783319434889;9783319434872

In humans, repeated exposure to the effects of events can lead to anticipation of these effects. This behaviour has been observed in infants from as young as 3months old. During infant experiments, the infants have been observed to predict either by pre-saccadic movements or reach actions according to the expected future outcome of the event. Event anticipation or prediction is necessary for such behaviours. In this paper we demonstrate prediction of object motion events using the adaptive learning tool Dev-PSchema. Results shows that the system is able to predict the linear motion outcome of the visual event using generalised schemas.

关键词： developmental robotics Psychologically inspired Action prediction

来源：评论

学校读者我要写书评

暂无评论

From Manipulation to Communicative Gesture

From Manipulation to Communicative Gesture

引用

5th ACM/IEEE International Conference on Human-Robot Interaction (HRI)

作者： Ou, Shichao Grupen, Rod Univ Massachusetts Lab Perceptual Robot Dept Comp Sci Amherst MA 01003 USA

ISBN: (纸本)9781424448937

This paper advocates an approach for learning communicative actions and manual skills in the same framework. We exploit a fundamental relationship between the structure of motor skills, intention, and communication. Communicative actions are acquired using the same learning framework and the same primitive states and actions that the robot uses to construct manual behavior for interacting with other objects in the environment. A prospective behavior algorithm is used to acquire modular policies for conveying intention and goals to nearby human beings and recruiting their assistance. The learning framework and a preliminary case study are presented in which a humanoid robot learns expressive communicative behavior incrementally by discovering the manual affordances of human beings. Results from interactions with 16 people provide support for the hypothesized benefits of this approach. Behavior reuse makes learning from relatively few interactions possible. This approach compliments other efforts in the field by grounding social behavior, and proposes a mechanism for negotiating a communicative vocabulary between humans and robots.

关键词： communicative behavior learning knowledge acquisition developmental robotics human-robot interaction

来源：评论

学校读者我要写书评

暂无评论

Continual Learning of Human-like Arm Postures

Continual Learning of Human-like Arm Postures

引用

IEEE International Conference on Development and Learning (ICDL)

作者： Gulletta, Gianpaolo Erlhagen, Wolfram Bicho, Estela Univ Minho Ctr Algoritmi Dept Ind Elect Guimaraes Portugal Univ Minho Ctr Math Dept Math Guimaraes Portugal

ISBN: (纸本)9781728162423

Inspired from established human motor control theories, our HUMP algorithm plans upper-limb collisions-free movements for anthropomorphic systems, which show kinematic human-like features [1]. Related cognitive issues can be further resolved when robots act as they are familiar with their workspace and can take initiative faster than in the early onsets of a task. Here, a continual learning technique is proposed to improve the performance of the HUMP under uncertainties of the items in a given scenario. Given the locality of the optimization-based HUMP algorithm, a meaningful initial guess, predicted from similar past motion experiences, can significantly reduce the computational cost and put the robot into action arguably faster than in the first attempts of planning with inexperienced initial guesses. This prediction is proposed to be incrementally refined by an optimal locally weighted regression method that operates on datasets of situational features that are regularly updated as new movements are planned by the robot in similar scenarios. The proposed cyclic experiential learner is tested on the selection of optimal human-like target postures in a reaching task with a large obstacle obstructing the straight-line path towards a given target. Results demonstrate the capability of extracting meaningful situational features in few sessions of online learning with a very limited size of the datasets. Comparisons with simple Euclidean locally weighted regression and random initializations showed the capability of planning target configurations of better quality with less computational cost. The proposed approach also exhibits to be robust against the interferences of new incoming samples depicting slightly changed situations of the same task.

关键词： developmental robotics continual learning cognitive robotics human-like arm posture prediction warm-starting IPOPT

来源：评论

学校读者我要写书评

暂无评论

Developing hierarchical anticipations via neural network-based event segmentation

Developing hierarchical anticipations via neural network-bas...

引用

IEEE International Conference on Development and Learning (ICDL)

作者： Gumbsch, Christian Adam, Maurits Elsner, Birgit Martius, Georg Butz, Martin, V Univ Tubingen Neuro Cognit Modeling Grp Tubingen Germany Max Planck Inst Intelligent Syst Autonomous Learning Grp Tubingen Germany Univ Potsdam Dev Psychol Potsdam Germany

ISBN: (数字)9781665413114

ISBN: (纸本)9781665413114

Humans can make predictions on various time scales and hierarchical levels. Thereby, the learning of event encodings seems to play a crucial role. In this work we model the development of hierarchical predictions via autonomously learned latent event codes. We present a hierarchical recurrent neural network architecture, whose inductive learning biases foster the development of sparsely changing latent state that compress sensorimotor sequences. A higher level network learns to predict the situations in which the latent states tend to change. Using a simulated robotic manipulator, we demonstrate that the system (i) learns latent states that accurately reflect the event structure of the data, (ii) develops meaningful temporal abstract predictions on the higher level, and (iii) generates goal-anticipatory behavior similar to gaze behavior found in eye-tracking studies with infants. The architecture offers a step towards the autonomous learning of compressed hierarchical encodings of gathered experiences and the exploitation of these encodings to generate adaptive behavior.

关键词： developmental robotics goal-predictive gaze event cognition temporal abstractions

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：