By Topic

Development and Learning and Epigenetic Robotics (ICDL), 2012 IEEE International Conference on

Date 7-9 Nov. 2012

Filter Results

Displaying Results 1 - 25 of 106
  • [Title page]

    Page(s): 1
    Save to Project icon | Request Permissions | PDF file iconPDF (69 KB)  
    Freely Available from IEEE
  • [Copyright notice]

    Page(s): 1
    Save to Project icon | Request Permissions | PDF file iconPDF (53 KB)  
    Freely Available from IEEE
  • [Title page]

    Page(s): 1
    Save to Project icon | Request Permissions | PDF file iconPDF (369 KB)  
    Freely Available from IEEE
  • Paper index

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | PDF file iconPDF (149 KB)  
    Freely Available from IEEE
  • A face recognition system that simulates perception impairments of autistic children

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1018 KB) |  | HTML iconHTML  

    We use a face recognition algorithm to model differences in perception between autistic and non autistic children. With our model it is possible to reproduce several phenomena of autism by assuming that autistic children lack the ability to abstract from horizontal invariants. In particular, we can explain why autistic children are able to better recognize faces from parts of the face while the overall recognition of faces is worse than in non autistic children. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Embodiment guides motor and spinal circuit development in vertebrate embryo and fetus

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2916 KB) |  | HTML iconHTML  

    We investigated whether motor and spinal circuit development in vertebrates can be accounted for by the properties underlying embodiment. We ran computer simulations of zebrafish embryo, canine and human fetus models with biologically plausible musculoskeletal bodies and spinal neural network, and quantitatively characterized their embodiments and movements by analyzing inter-muscle connectivities. In computer simulations in the human and canine fetus models, we found that development of the embodiment causes changes in movements and increases their complexity, corresponding to the same manner as mammalian motor development. Further, we showed that interaction with the environment as structured by the embodiment can drive the self-organization of the spinal circuit and trigger important developmental motor transitions, which engender coordinated side-to-side alternating movements in the zebrafish embryo model and left-right alternation of the legs in the human fetus model. Our results suggest that embodiment possesses a multitude of regularities that can guide early development. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Robotic model of the contribution of gesture to learning to count

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (422 KB) |  | HTML iconHTML  

    In this paper a robotic connectionist model of the contribution of gesture to learning to count is presented. By formulating a recurrent artificial neural network model of the phenomenon and assessing its performance without and with gesture it is demonstrated that the proprioceptive signal connected with gesture carries information which may be exploited when learning to count. The behaviour of the model is similar to that of human children in terms of the effect of gesture and the size of the counted set, although the detailed patterns of errors made by the model and human children are different. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Emergence of color constancy illusion through reinforcement learning with a neural network

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1619 KB) |  | HTML iconHTML  

    Our parallel and flexible brain that must be the origin of our flexibility processes visual signals without being noticed, and due to the unawareness, the contradiction between our perception after the process and original visual property is exposed as “Optical Illusion”. The authors form the hypothesis that optical illusion can be acquired through or supported by the learning so as that we behave more appropriately in everyday life. In this paper, “color constancy” is focused on and the authors try to explain its emergence through the learning of a simple “colored-object guidance” task by reinforcement learning with a neural network whose inputs are raw image signals. In the task, it is required to move an object whose color is chosen randomly to the proper location that differs depending on the object color. Half of the field is covered by a translucent filter whose color and angle are chosen randomly at each episode. It was observed that the hidden neurons came to represent the object color mainly not depending on the filter color after reinforcement learning. In the subsequent supervised learning and test, the neural network with new output neurons was trained to output the object color only under the condition of no filter, but, when images covered by colored filter were the input as test patterns after learning, the network outputs were very close to the original object color. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Generation of condition-dependent reaching movements based on layered associative networks

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (439 KB) |  | HTML iconHTML  

    This paper proposes a hierarchical network that enables an information processing nesting structure for tool-use that represents a relationship between a hand end-effector and a tool target when reaching for the tool, and an additional relationship between the tip of the tool end-effector and an object target after taking the tool. The network consists of associative networks whose lower layer networks associate the multimodal information under various conditions. The higher layer network associates the temporal order of the lower network states. A simulation experiment shows that the proposed network successfully generates reaching movements regardless of the conditions. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Computing affect in autonomous agents

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (304 KB) |  | HTML iconHTML  

    In previous work we proposed that affect can be modeled in an autonomous agent. We report progress made since then-specifically, an improved understanding of the basis of our approach, a new version of the model used earlier, rationalized both in terms of the underlying neuroscience and the equations used to compute affect. We describe plans for future work to remove constants that we presently have hard coded into the model. We also report some of the difficulties we face: the need for a stronger empirical basis for the equations we use, and the difficulty in identifying success criteria. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Curiosity-driven phonetic learning

    Page(s): 1 - 8
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2108 KB) |  | HTML iconHTML  

    This article studies how developmental phonetic learning can be guided by pure curiosity-driven exploration, also called intrinsically motivated exploration. Phonetic learning refers here to learning how to control a vocal tract to reach acoustic goals. We compare three different exploration strategies for learning the auditory-motor inverse model: random motor exploration, random goal selection with reaching, and curiosity-driven active goal selection with reaching. Using a realistic vocal tract model, we show how intrinsically motivated learning driven by competence progress can generate automatically developmental structure in both articulatory and auditory modalities, displaying patterns in line with some experimental data from infants. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Understanding the development of motion processing by characterizing optic flow experienced by infants and their mothers

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (447 KB) |  | HTML iconHTML  

    Understanding the development of mature motion processing may require knowledge about the statistics of the visual input that infants are exposed to, how these change across development, and how they influence the maturation of motion-sensitive brain networks. Here we develop a set of techniques to study the optic flow experienced by infants and mothers during locomotion as a first step toward a broader analysis of the statistics of the natural visual environment during development. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Simultaneous concept formation driven by predictability

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (506 KB) |  | HTML iconHTML  

    This study is conducted in the context of developmental learning in embodied agents who have multiple data sources (sensors) at their disposal. We describe an online learning method that simultaneously discovers “meaningful” concepts in the associated processing streams, extending methods such as PCA, SOM or sparse coding to the multimodal case. In addition to the avoidance of redundancies in the concepts derived from single modalities, we claim that “meaningful” concepts are those who have statistical relations across modalities. This is a reasonable claim because measurements by different sensors often have common cause in the external world and therefore carry correlated information. To capture such cross-modal relations while avoiding redundancy of concepts, we propose a set of interacting self-organization processes which are modulated by local predictability. To validate the fundamental applicability of the method, we conduct a plausible simulation experiment with synthetic data and find that those concepts which are predictable from other modalities successively “grow”, i.e., become over-represented, whereas concepts that are not predictable become systematically under-represented. We conclude the article by a discussion of applicability in real-world robotics scenarios. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Emergent proximo-distal maturation through adaptive exploration

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1049 KB) |  | HTML iconHTML  

    Life-long robot learning in the high-dimensional real world requires guided and structured exploration mechanisms. In this developmental context, we investigate here the use of the recently proposed PICMAES2 episodic reinforcement learning algorithm, which is able to learn high-dimensional motor tasks through adaptive control of exploration. By studying PICMAES2 in a reaching task on a simulated arm, we observe two developmental properties. First, we show how PICMAES2 autonomously and continuously tunes the global exploration/exploitation tradeoff, allowing it to re-adapt to changing tasks. Second, we show how PICMAES2 spontaneously self-organizes a maturational structure whilst exploring the degrees-of-freedom (DOFs) of the motor space. In particular, it automatically demonstrates the so-called proximo-distal maturation observed in humans: after first freezing distal DOFs while exploring predominantly the most proximal DOF, it progressively frees exploration in DOFs along the proximo-distal body axis. These emergent properties suggest the use of PICMAES2 as a general tool for studying reinforcement learning of skills in life-long developmental learning contexts. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Artificial aesthetic: An interesting framework for epigenetic robotics

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (371 KB) |  | HTML iconHTML  

    We propose here a first experiment, where museum visitors were asked to educate the aesthetic preferences of a robot according to their own aesthetic preferences. This provides an interesting framework both for testing developmental models and studying human-robot interactions. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Robot acquisition of lexical meaning - moving towards the two-word stage

    Page(s): 1 - 7
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (269 KB) |  | HTML iconHTML  

    We report on experiments and analyses dealing with the acquisition of lexical meaning in which prosodic analysis and extraction of salient words are associated with a robots sensorimotor perceptions in an attempt to ground these words in the robots own embodied sensorimotor experience. We focus here on three key areas, the selection of salient words based on prosodic clues, expression of words by the robot at a two-word stage to reflect learning and grammatically correct presentation, and an in-depth analysis of the relationship between words and the robots sensorimotor perceptions. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Incremental learning of an optical flow model for sensorimotor anticipation in a mobile robot

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (130 KB) |  | HTML iconHTML  

    In this paper we study the mechanisms that enable an active agent to make long-term predictions of optical flow with a model that is learned dynamically. We analyse the optical flow distribution in terms of space and time, that is, what are the experienced optical flow values and how do they change in time. We show how complex the posterior distributions become when long-term predictions are needed, which breaks time-consistency assumption. The choice of one predictor or another should be made in terms of how the data is distributed. Moreover, we use a generic state-of-the-art incremental online learning algorithm [8] for the task of building a model to predict the optical flow perceived by a mobile robot. Finally, as an application, the model is also used to learn a simple predictor for anticipating an imminent collision. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Intrinsically motivated model learning for a developing curious agent

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (146 KB) |  | HTML iconHTML  

    Reinforcement Learning (RL) agents are typically deployed to learn a specific, concrete task based on a pre-defined reward function. However, in some cases an agent may be able to gain experience in the domain prior to being given a task. In such cases, intrinsic motivation can be used to enable the agent to learn a useful model of the environment that is likely to help it learn its eventual tasks more efficiently. This paper presents the TEXPLORE with Variance-And-Novelty-Intrinsic-Rewards algorithm (TEXPLORE-VANIR), an intrinsically motivated model-based RL algorithm. The algorithm learns models of the transition dynamics of a domain using random forests. It calculates two different intrinsic motivations from this model: one to explore where the model is uncertain, and one to acquire novel experiences that the model has not yet been trained on. This paper presents experiments demonstrating that the combination of these two intrinsic rewards enables the algorithm to learn an accurate model of a domain with no external rewards and that the learned model can be used afterward to perform tasks in the domain. While learning the model, the agent explores the domain in a developing and curious way, progressively learning more complex skills. In addition, the experiments show that combining the agent's intrinsic rewards with external task rewards enables the agent to learn faster than using external rewards alone. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Rapidly learning preconditions for means-ends behaviour using active learning

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (610 KB) |  | HTML iconHTML  

    In [1], we argue that ongoing development may be the result of a set of developmental mechanisms which are in continuous operation during infancy. One such mechanism identified is sensorimotor differentiation. Sensorimotor differentiation allows infants to generate new behaviours by modifying old ones. For example a young infant has a behaviour for waving an object back and forth on a table surface. At some later point, this behaviour becomes differentiated to produce a behaviour for deliberately displacing an object to one side in order to retrieve a visible toy behind it (see Figure I). View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An adaptive neuro-fuzzy approach for semantic analysis of broadcast soccer video

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (278 KB) |  | HTML iconHTML  

    This paper presents an approach for automatic annotation of soccer video based on the semantic events occurred inside it. The goal of this paper is to propose a flexible system that can be able to be used with minimum reliance on predefined patterns in event detection process. To achieve this goal, we propose a fuzzy inference system (FIS) implemented in the framework of an adaptive neural network which combines the self-learning capability of neural networks with explicit knowledge representation and precision of fuzzy based classification systems. This method provides the capability for fuzzy systems to learn information about a set of data in order to determine the parameters of membership functions (MFs) automatically and generate a set of fuzzy rules that best allow the FIS to track the input/output data. The proposed method is multimodal and employs statistical information from a set of audiovisual features that are organized in a hierarchical structure as input and produces semantic concepts corresponding to the occurred events. Experimental results conducted on a large set of soccer videos demonstrate the effectiveness of the proposed approach. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An infant inspired model of reaching for a humanoid robot

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (241 KB) |  | HTML iconHTML  

    Infants demonstrate remarkable talents in learning to control their sensor and motor systems. In particular the ability to reach to objects using visual feedback requires overcoming several issues related to coordination, spatial transformations, redundancy, and complex learning spaces, that are also challenges for robotics. The development sequence from tabula rasa to early successful reaching includes learning of saccade control, gaze control, torso control, and visually elicited reaching and grasping in 3D space. This sequence is an essential progression in the acquisition of manipulation behaviour. In this paper we outline the biological and psychological processes behind this sequence, and describe how they can be interpreted to enable cumulative learning of reaching behaviours in robots. Our implementation on an iCub robot produces reaching and manipulation behaviours from scratch in around 2.5 hours. We show snapshots of the learning spaces during this process, and comment on how timing of stage transition impacts on learning. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improvement proposals to intrinsically motivational robotics

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (106 KB) |  | HTML iconHTML  

    In this paper, we first give a brief background on artificial curiosity and intrinsic motivation as it is studied in the developmental robotics research community. We then introduce some theoretical improvements for mechanisms related to intrinsically motivational living algorithms as introduced by SAGG-RIAC algorithm. Finally we conclude by drawing the way for our future experimental results using these new improvements. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The strategic student approach for life-long exploration and learning

    Page(s): 1 - 8
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (328 KB) |  | HTML iconHTML  

    This article introduces the strategic student metaphor: a student has to learn a number of topics (or tasks) to maximize its mean score, and has to choose strategically how to allocate its time among the topics and/or which learning method to use for a given topic. We show that under which conditions a strategy where time allocation or learning method is chosen from the easier to the more complex topic is optimal. Then, we show an algorithm, based on multi-armed bandit techniques, that allows empirical online evaluation of learning progress and approximates the optimal solution under more general conditions. Finally, we show that the strategic student problem formulation allows to view in a common framework many previous approaches to active and developmental learning. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multi-optima exploration with adaptive Gaussian mixture model

    Page(s): 1 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1830 KB) |  | HTML iconHTML  

    In learning by exploration problems such as reinforcement learning (RL), direct policy search, stochastic optimization or evolutionary computation, the goal of an agent is to maximize some form of reward function (or minimize a cost function). Often, these algorithms are designed to find a single policy solution. We address the problem of representing the space of control policy solutions by considering exploration as a density estimation problem. Such representation provides additional information such as shape and curvature of local peaks that can be exploited to analyze the discovered solutions and guide the exploration. We show that the search process can easily be generalized to multi-peaked distributions by employing a Gaussian mixture model (GMM) with an adaptive number of components. The GMM has a dual role: representing the space of possible control policies, and guiding the exploration of new policies. A variation of expectation-maximization (EM) applied to reward-weighted policy parameters is presented to model the space of possible solutions, as if this space was a probability distribution. The approach is tested in a dart game experiment formulated as a black-box optimization problem, where the agent's throwing capability increases while it chases for the best strategy to play the game. This experiment is used to study how the proposed approach can exploit new promising solution alternatives in the search process, when the optimality criterion slowly drifts over time. The results show that the proposed multi-optima search approach can anticipate such changes by exploiting promising candidates to smoothly adapt to the change of global optimum. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Socially guided intrinsically motivated learner

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1902 KB) |  | HTML iconHTML  

    This paper studies the coupling of two learning strategies: internally guided learning and social interaction. We present Socially Guided Intrinsic Motivation by Demonstration (SGIM-D) and its interactive learner version Socially Guided Intrinsic Motivation with Interactive learning at the Meta level (SGIM-IM), which are algorithms for learning inverse models in high dimensional continuous sensorimotor spaces. After describing the general framework of our algorithms, we illustrate with a fishing experiment. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.