Developmental Cognitive Robotics

For decades, the world has thrown billions at "artificial intelligence," believing the answer was always more data, more compute, more prediction. I'm here to say: it isn't. Today's "AI" is a sophisticated exercise in statistical mimicry, what I call Artificial Prediction (AP), not intelligence, not mind, not anything remotely alive. What we're missing, and what the field refuses to admit, is the root of all real intelligence: consciousness, memory, development, and feeling. My manifesto is not just critique. It's a blueprint, a demand, and a reckoning. I lay out the true science of intelligence as seen in trees, dogs, and humans; dissect the failures and empty promises of AP; and map the only credible route to synthetic intelligence, not artificial, not imitation, but a new branch of mind as real as anything that's ever walked, grown, or thought on this earth. It's time to retrain the entire industry. Here's the foundation.

Evidence from developmental as well as neuroscientific studies suggest that finger counting activity plays an important role in the acquisition of numerical skills in children. It has been claimed that this skill helps in building motor-based representations of number that continue to influence number processing well into adulthood, facilitating the emergence of number concepts from sensorimotor experience through a bottom-up process. The act of counting also involves the acquisition and use of a verbal number system of which number words are the basic building blocks. Using a Cognitive Developmental Robotics paradigm we present results of a modeling experiment on whether finger counting and the association of number words (or tags) to fingers, could serve to bootstrap the representation of number in a cognitive robot, enabling it to perform basic numerical operations such as addition. The cognitive architecture of the robot is based on artificial neural networks, which enable the robot to learn both sensorimotor skills (finger counting) and linguistic skills (using number words). The results obtained in our experiments show that learning the number words in sequence along with finger configurations helps the fast building of the initial representation of number in the robot. Number knowledge, is instead, not as efficiently developed when number words are learned out of sequence without finger counting. Furthermore, the internal representations of the finger configurations themselves, developed by the robot as a result of the experiments, sustain the execution of basic arithmetic operations, something consistent with evidence coming from developmental research with children. The model and experiments demonstrate the importance of sensorimotor skill learning in robots for the acquisition of abstract knowledge such as numbers.

D 1.5 Cognitive organisation for sustaining eSMCs A study of the different organisational requirements to acquire, sustain and negotiate several eSMCs in a same cognitive agent. Deliverable specification The deliverable is linked to Task 1.3: Requirements for acquiring, sustaining and integrating eSMCs (UPV, UOS). An examination will be carried out regarding the plasticity mechanisms that allow embodied systems to successfully acquire new eSMCs or adapt to disruptions in sensors, effectors or other aspects of body morphology. Principles of regulation of plasticity will be studied (supported by modelling work in WP 2) to generate neural models that deal with acquisition, resolution of interference and higher-level combination of eSMCs. Particular attention will be paid to situations of conflict between eSMCs, for instance, in objects with multiple affordances. We will investigate whether there is a need for dedicated (neural) mechanisms at play during the acquisition, the adaptive maintenance and the integration of eSMCs, or whether, in contrast, a single organisational principle could explain the three phenomena. Following results from Task 1.1 on the distinction between the requirements for minimal action and minimal SMC-based skills, we will examine organisational support for the requirements of repeatability, reliability and plastic incorporation of eSMCs and investigate to what extent they might support the single organisational principle hypothesis. Expected results and presentation/verification: Neural model on plasticity mechanisms relevant for acquisition, resolution of interference and higher-level combination of eSMCs.

This paper presents an interdisciplinary study of the role of the head in multisensory integration and motor-control organization for the production of voluntary spatial actions. It combines elements from biology and engineering. First, morphological and behavioral characteristics of animals able to perform voluntary spatial actions through evolution are examined. The complexity of state-space representation and observation of multi-joint mobile robots is then described in the context of automatic control, and perception-related characteristics brought by the presence of a head are presented from the perspective of signal processing. Finally, the role of the head in locomotion and manipulation for animals and robots is discussed, paving the way for future robot design.

Recent work in Human-Robot Interaction (HRI) investigates the role of human users as teachers from which robots can flexibly learn new personalised skills through interaction. However, existing human-robot teaching methods remain largely unintuitive for the end user and require significant effort to adapt to the way the robot learns. This paper envisions the use of dog training methods as a starting point for HRI researchers to develop more intuitive interactions between human teachers and robot learners. We provide a design framework (called FETCH-R) aimed at guiding the conception of interactions between human teachers and robot learners inspired by dog training. This work paves the way towards the use of animal training as an inspiration to create human-robot teaching protocols that promote engagement, ease-of-use, and fosters human-robot relationships.

In this paper, we propose a formal definition of the perception as a behavioral dynamical attraction basin. The perception is built from the integration of the sensori-motor flow. Psychological considerations and robotic experiments on an embodied "intelligent" system are provided to show how this definition can satisfy both psychologist and robotician point of view.

Traditionally, in robotics, artificial intelligence and neuroscience, there has been a focus on the study of the control or the neural system itself. Recently there has been an increasing interest in the notion of embodiment not only in robotics and artificial intelligence, but also in the neurosciences, psychology and philosophy. In this paper, we introduce the notion of morphological computation, and demonstrate how it can be exploited on the one hand for designing intelligent, adaptive robotic systems, and on the other hand for understanding natural systems. While embodiment has often been used in its trivial meaning, i.e. bintelligence requires a bodyQ, the concept has deeper and more important implications, concerned with the relation between physical and information (neural, control) processes. Morphological computation is about connecting body, brain and environment. A number of case studies are presented to illustrate the concept. We conclude with some speculations about potential lessons for neuroscience and robotics.

We present a system to learn task representations from ambiguous feedback. We consider an inverse reinforcement learner that receives feedback from a teacher with an unknown and noisy protocol. The system needs to estimate simultaneously what the task is (i.e. how to find a compact representation to the task goal), and how the teacher is providing the feedback. We further explore the problem of ambiguous protocols by considering that the words used by the teacher have an unknown relation with the action and meaning expected by the robot. This allows the system to start with a set of known signs and learn the meaning of new ones. We present computational results that show that it is possible to learn the task under a noisy and ambiguous feedback. Using an active learning approach, the system is able to reduce the length of the training period.

This paper presents an algorithm to bootstrap shared understanding in a human-robot interaction scenario where the user teaches a robot a new task using teaching instructions yet unknown to it. In such cases, the robot needs to estimate simultaneously what the task is and the associated meaning of instructions received from the user. For this work, we consider a scenario where a human teacher uses initially unknown spoken words, whose associated unknown meaning is either a feedback (good/bad) or a guidance (go left, right, ...). We present computational results, within an inverse reinforcement learning framework, showing that a) it is possible to learn the meaning of unknown and noisy teaching instructions, as well as a new task at the same time, b) it is possible to reuse the acquired knowledge about instructions for learning new tasks, and c) even if the robot initially knows some of the instructions' meanings, the use of extra unknown teaching instructions improves learning efficiency.

Rodney Brooks (1991) put forth the idea that during an agent's interaction with its environment, representations of the world often stand in the way. Instead, using the world as its own best model, i.e. interacting with it directly without making models, often leads to better and more natural behavior. The same perspective can be applied to representations of the agent's body. I analyze different examples from biology-octopus and humans in particular-and compare them with robots and their body models. At one end of the spectrum, the octopus, a highly intelligent animal, largely relies on the mechanical properties of its arms and peripheral nervous system. No central representations or maps of its body were found in its central nervous system. Primate brains do contain areas dedicated to processing body-related information and different body maps were found. Yet, these representations are still largely implicit and distributed and some functionality is also offloaded to the periphery. Robots, on the other hand, rely almost exclusively on their body models when planning and executing behaviors. I analyze the pros and cons of these different approaches and propose what may be the best solution for robots of the future.

Humans and animals excel in combining information from multiple sensory modalities, controlling their complex bodies, adapting to growth, failures, or using tools. These capabilities are also highly desirable in robots. They are displayed by machines to some extent - yet, as is so often the case, the artificial creatures are lagging behind. The key foundation is an internal representation of the body that the agent - human, animal, or robot - has developed. In the biological realm, evidence has been accumulated by diverse disciplines giving rise to the concepts of body image, body schema, and others. In robotics, a model of the robot is an indispensable component that enables to control the machine. In this article I compare the character of body representations in biology with their robotic counterparts and relate that to the differences in performance that we observe. I put forth a number of axes regarding the nature of such body models: fixed vs. plastic, amodal vs. modal, explic...

One of the possible benefits of robot-mediated education is the effect of the robot becoming a catalyst between people and facilitating learning. In this study, the authors focused on an asynchronous active learning method mediated by robots. Active learning is believed to help students continue learning and develop the ability to think independently. Therefore, the authors improved the UGA (User Generated Agent) system that we have created for long-term active learning in COVID-19 to create an environment where children introduce books to each other via robots. The authors installed the robot in an elementary school and conducted an experiment lasting more than a year. As a result, it was confirmed that the robot could continue to be used without getting bored even over a long period of time. They also analyzed how the children created the contents by analyzing the contents that had a particularly high number of views. In particular, the authors observed changes in children’s behav...

Evidence from developmental as well as neuroscientific studies suggest that finger counting activity plays an important role in the acquisition of numerical skills in children. It has been claimed that this skill helps in building motor-based representations of number that continue to influence number processing well into adulthood, facilitating the emergence of number concepts from sensorimotor experience through a bottom-up process. The act of counting also involves the acquisition and use of a verbal number system of which number words are the basic building blocks. Using a Cognitive Developmental Robotics paradigm we present results of a modeling experiment on whether finger counting and the association of number words (or tags) to fingers, could serve to bootstrap the representation of number in a cognitive robot, enabling it to perform basic numerical operations such as addition. The cognitive architecture of the robot is based on artificial neural networks, which enable the robot to learn both sensorimotor skills (finger counting) and linguistic skills (using number words). The results obtained in our experiments show that learning the number words in sequence along with finger configurations helps the fast building of the initial representation of number in the robot. Number knowledge, is instead, not as efficiently developed when number words are learned out of sequence without finger counting. Furthermore, the internal representations of the finger configurations themselves, developed by the robot as a result of the experiments, sustain the execution of basic arithmetic operations, something consistent with evidence coming from developmental research with children. The model and experiments demonstrate the importance of sensorimotor skill learning in robots for the acquisition of abstract knowledge such as numbers.

Humans and animals excel in combining information from multiple sensory modalities, controlling their complex bodies, adapting to growth or failures, or using tools. The key foundation is an internal representation of the body that the agent—human, animal, or robot—has developed. In the biological realm, evidence has been accumulating in diverse disciplines, giving rise to the concepts of body image, body schema, and others. In robotics, a model of the robot is an indispensable component that enables to control the machine. This chapter compares the character of body representations in biology with their robotic counterparts and relates that to the differences in performance observed. Conclusions are drawn about how robots can inform the biological sciences dealing with body representations and which of the features of the ‘body in the brain’ should be transferred to robots, giving rise to more adaptive and resilient self-calibrating machines.

Somatosensory inputs can be grossly divided into tactile (or cutaneous) and proprioceptive -- the former conveying information about skin stimulation, the latter about limb position and movement. The principal proprioceptors are constituted by muscle spindles, which deliver information about muscle length and speed. In primates, this information is relayed to the primary somatosensory cortex and eventually the posterior parietal cortex, where integrated information about body posture (postural schema) is presumably available. However, coming from robotics and seeking a biologically motivated model that could be used in a humanoid robot, we faced a number of difficulties. First, it is not clear what neurons in the ascending pathway and primary somatosensory cortex code. To an engineer, joint angles would seem the most useful variables. However, the lengths of individual muscles have nonlinear relationships with the angles at joints. Kim et al. (Neuron, 2015) found different types of ...

A large body of compelling evidence has been accumulated demonstrating that embodiment—the agent’s physical setup, including its shape, materials, sensors, and actuators—is constitutive for any form of cognition, and, as a consequence, models of cognition need to be embodied. In contrast to methods from empirical sciences to study cognition, robots can be freely manipulated and virtually all key variables of their embodiment and control programs can be systematically varied. As such, they provide an extremely powerful tool of investigation. We present a robotic, bottom-up, or developmental approach, focusing on three stages: (1) low-level behaviors like walking and reflexes, (2) learning regularities in sensorimotor spaces, and (3) human-like cognition. We also show that robotic-based research is not only a productive path to deepening our understanding of cognition, but that robots can strongly benefit from human-like cognition in order to become more autonomous, robust, resilient,...

Using the iCub humanoid robot with an artificial pressure-sensitive skin, we investigate how representations of the whole skin surface resembling those found in primate primary somatosensory cortex can be formed from local tactile stimulations traversing the body of the physical robot. We employ the well-known self-organizing map (SOM) algorithm and introduce its modification that makes it possible to restrict the maximum receptive field (MRF) size of neuron groups at the output layer. This is motivated by findings from biology where basic somatotopy of the cortical sheet seems to be prescribed genetically and connections are localized to particular regions. We explore different settings of the MRF and the effect of activity-independent (input-output connections constraints implemented by MRF) and activity-dependent (learning from skin stimulations) mechanisms on the formation of the tactile map. The framework conveniently allows one to specify prior knowledge regarding the skin topology and thus to effectively seed a particular representation that training shapes further. Furthermore, we show that the MRF modification facilitates learning in situations when concurrent stimulation at non-adjacent places occurs ("multi-touch"). The procedure was sufficiently robust and not intensive on the data collection and can be applied to any robots where representation of their "skin" is desirable.

Animals and humans engage in an enormous variety of behaviors which are orchestrated through a complex interaction of physical and informational processes: The physical interaction of the bodies with the environment is intimately coupled with informational processes in the animal's brain. A crucial step toward the mastery of all these behaviors seems to be to understand the flows of information in the sensorimotor networks. In this study, we have performed a quantitative analysis in an artificial agent — a running quadruped robot with multiple sensory modalities — using tools from information theory (transfer entropy). Starting from very little prior knowledge, through systematic variation of control signals and environment, we show how the agent can discover the structure of its sensorimotor space, identify proprioceptive and exteroceptive sensory modalities, and acquire a primitive body schema. In summary, we show how the analysis of directed information flows in an agent'...

Learning to perceive is faced with a classical paradox: if understanding is required for perception, how can we learn to perceive something new, something we do not yet understand? According to the sensorimotor approach, perception involves mastery of regular sensorimotor co-variations that depend on the agent and the environment, also known as the "laws" of sensorimotor contingencies (SMCs). In this sense, perception involves enacting relevant sensorimotor skills in each situation. It is important for this proposal that such skills can be learned and refined with experience and yet up to this date, the sensorimotor approach has had no explicit theory of perceptual learning. The situation is made more complex if we acknowledge the open-ended nature of human learning. In this paper we propose Piaget's theory of equilibration as a potential candidate to fulfill this role. This theory highlights the importance of intrinsic sensorimotor norms, in terms of the closure of sensorimotor schemes. It also explains how the equilibration of a sensorimotor organization faced with novelty or breakdowns proceeds by reshaping pre-existing structures in coupling with dynamical regularities of the world. This way learning to perceive is guided by the equilibration of emerging forms of skillful coping with the world. We demonstrate the compatibility between Piaget's theory and the sensorimotor approach by providing a dynamical formalization of equilibration to give an explicit micro-genetic account of sensorimotor learning and, by extension, of how we learn to perceive. This allows us to draw important lessons in the form of general principles for open-ended sensorimotor learning, including the need for an intrinsic normative evaluation by the agent itself. We also explore implications of our micro-genetic account at the personal level.

Embodied Cognitive Robotics focuses its attention on the design of artificial agents capable of performing cognitive tasks autonomously. A central issue in this consists in studying process by which agents learn through interaction with their environment. Embodied Cognitive Robotics aims to implement models of cognitive processes coming from Cognitive Sciences. The guidelines in this research area are a direct response to the shortcomings of Classical Artificial Intelligence, where high-level tasks and behaviors were studied. This article describes the work carried out in the Cognitive Robotics Laboratory at the Universidad Autónoma del Estado de Morelos (UAEM). Our work is based on the concept of low-level sensorimotor schemes coded by Internal Models, thus falling as a matter of course within the tenets of Embodied Cognition, particularly with the idea that cognition must be understood as occurring in agents that have a body with which they interact in a specific environment. It i...

Deep unsupervised learning in stochastic recurrent neural networks with many layers of hidden units is a recent breakthrough in neural computation research. These networks build a hierarchy of progressively more complex distributed representations of the sensory data by fitting a hierarchical generative model. In this article we discuss the theoretical foundations of this approach and we review key issues related to training, testing and analysis of deep networks for modeling language and cognitive processing. The classic letter and word perception problem of McClelland and Rumelhart (1981) is used as a tutorial example to illustrate how structured and abstract representations may emerge from deep generative learning. We argue that the focus on deep architectures and generative (rather than discriminative) learning represents a crucial step forward for the connectionist modeling enterprise, because it offers a more plausible model of cortical learning as well as a way to bridge the gap between emergentist connectionist models and structured Bayesian models of cognition.

A self-organizing neural network is described that can associate between different modalities and also has the ability to learn perceptual sequences. This architecture is a step towards the development of a complete agent containing simplified versions of all major neural subsystems in a mammal. It aims at exploring as well as takes inspiration from the idea that cognitive function involves an internal simulation of perception and movement. We have tested the architecture in simulations as well as together with real sensors with very encouraging results. Perceptul activity in the brain is normally elicited from the sense organs, but it has recently been argued that a crucial aspect of cognitive function in biological organisms is the ability to simulate perception, that is, to elicit perceptual activity from other brain areas rather than via sensory afferents (Hesslow, 2002; Grush, 2004). Among other things, perceptual simulation could explain the appearance of an inner world (Hesslow and Jirenhed, 2007) and also how an organism could try out various courses of action 'in the mind'. One plausible source of simulated perceptual activity might be activity in another sensory modality. A familiar example is when the texture of an object that is felt in the pocket can evoke visual images/expectations of the object. A more dramatic biological illustration is the McGurk-MacDonald effect. If you hear a person making the sound /ba/ but the sound is superimposed on a video recording on which you do not see the lips closing, you may hear the sound /da/ instead (McGurk and MacDonald, 1976). Accordingly, a bio-inspired autonomous robot should be able to learn on its own and to automatically develop sensory representations of its different sensory modalities. These representations should also co-develop so that suitable activity in some modalities elicits appropriate activity in other sensory modalities as well. It would also be desirable with an ability to remember perceptual sequences, and that a corresponding sequence of activity evokes a suitable sequence of activity in the other modalities of the robot. With these traits a robot learning to navigate through an environment could remember a sequence of e.g. visual land

D 1.5 Cognitive organisation for sustaining eSMCs A study of the different organisational requirements to acquire, sustain and negotiate several eSMCs in a same cognitive agent. Deliverable specification The deliverable is linked to Task 1.3: Requirements for acquiring, sustaining and integrating eSMCs (UPV, UOS). An examination will be carried out regarding the plasticity mechanisms that allow embodied systems to successfully acquire new eSMCs or adapt to disruptions in sensors, effectors or other aspects of body morphology. Principles of regulation of plasticity will be studied (supported by modelling work in WP 2) to generate neural models that deal with acquisition, resolution of interference and higher-level combination of eSMCs. Particular attention will be paid to situations of conflict between eSMCs, for instance, in objects with multiple affordances. We will investigate whether there is a need for dedicated (neural) mechanisms at play during the acquisition, the adaptive maintenance and the integration of eSMCs, or whether, in contrast, a single organisational principle could explain the three phenomena. Following results from Task 1.1 on the distinction between the requirements for minimal action and minimal SMC-based skills, we will examine organisational support for the requirements of repeatability, reliability and plastic incorporation of eSMCs and investigate to what extent they might support the single organisational principle hypothesis. Expected results and presentation/verification: Neural model on plasticity mechanisms relevant for acquisition, resolution of interference and higher-level combination of eSMCs.

Developmental studies have suggested that infants' action is goal-directed. When imitating an action, younger infants tend to reproduce the goal while ignoring the means (i.e., the movement to achieve the goal) whereas older infants can imitate both. We suggest that the developmental dynamics of a Recurrent Neural Network with Parametric Bias (RNNPB) may explain the mechanism of infant development. Our RNNPB model was trained to reproduce six types of actions (2 different goals x 3 different means), during which parametric biases were self-organized to represent the difference with respect to both the goal and means. Our analysis of the self-organizing process of the parametric biases revealed an infant-like developmental change in action learning: the RNNPB first adapted to the goal and then to the means. The different saliency of these two features caused this phased development. We discuss the analogy of our result to infant action development.

Developmental studies have suggested that infants' action is goal-directed. When imitating an action, younger infants tend to reproduce the goal while ignoring the means (i.e., the movement to achieve the goal) whereas older infants can imitate both. We suggest that the developmental dynamics of a Recurrent Neural Network with Parametric Bias (RNNPB) may explain the mechanism of infant development. Our RNNPB model was trained to reproduce six types of actions (2 different goals x 3 different means), during which parametric biases were self-organized to represent the difference with respect to both the goal and means. Our analysis of the self-organizing process of the parametric biases revealed an infant-like developmental change in action learning: the RNNPB first adapted to the goal and then to the means. The different saliency of these two features caused this phased development. We discuss the analogy of our result to infant action development.

A self-organizing neural network is described that can associate between different modalities and also has the ability to learn perceptual sequences. This architecture is a step towards the development of a complete agent containing simplified versions of all major neural subsystems in a mammal. It aims at exploring as well as takes inspiration from the idea that cognitive function involves an internal simulation of perception and movement. We have tested the architecture in simulations as well as together with real sensors with very encouraging results.

The research that we have been involved in, and will continue to do, starts from the insight that in order to understand and design intelligent behavior, we must adopt an embodied perspective, i.e. we must take the entire agent, including its shape or morphology, the materials out of which it is built, and its interaction with the environment into account, in addition to the neural control. A lot of our research in the past has been on relatively low-level sensory-motor tasks such as locomotion (e.g. walking, running, jumping), navigation, and grasping. While this research is of interest in itself, in the context of artificial intelligence and cognitive science, this leads to the question of what these kinds of tasks have to do with higher levels of cognition, or to put it more provocatively, "What does walking have to do with thinking?" This question is of course reminiscent of the notorious "symbol grounding problem". In contrast to most of the research on symbol grounding, we propose to exploit the dynamic interaction between the embodied agent and the environment as the basis for grounding. We use the term "morphological computation" to designate the fact that some of the control or computation can be taken over by the dynamic interaction derived from morphological properties (e.g. the passive forward swing of the leg in walking, the spring-like properties of the muscles, and the weight distribution). By taking morphological computation into account, an agent will be able to achieve not only faster, more robust, and more energy-efficient behavior, but also more situated exploration by the agent for the comprehensive understanding of the environment.

In the field of developmental robotics, a lot of attention has been devoted to algorithms that allow agents to build up skills through sensorimotor interaction. Such interaction is largely affected by the agent's morphology, that is, its shape, limb articulation, as well as the position and density of sensors on its body surface. Despite its importance, the impact of morphology on behavior has not been systematically addressed. In this paper, we take inspiration from the human vision system, and demonstrate using a binocular active vision platform why sensor morphology in combination with other properties of the body, are essential conditions to achieve coordinated visual behavior (here, vergence). Specifically, to evaluate the effect of sensor morphology on behavior, we present an information-theoretic analysis quantifying the statistical regularities induced through sensorimotor interaction. Our results show that only for an adequate sensor morphology, vergence increases the amount of information structure in the sensorimotor loop.

Children typically learn basic numerical and arithmetic principles using finger-based representations. However, whether or not reliance on finger-based representations is beneficial or detrimental is the subject of an ongoing debate between researchers in neurocognition and mathematics education. From the neurocognitive perspective, finger counting provides multisensory input, which conveys both cardinal and ordinal aspects of numbers. Recent data indicate that children with good finger-based numerical representations show better arithmetic skills and that training finger gnosis, or "finger sense," enhances mathematical skills. Therefore neurocognitive researchers conclude that elaborate finger-based numerical representations are beneficial for later numerical development. However, research in mathematics education recommends fostering mentally based numerical representations so as to induce children to abandon finger counting. More precisely, mathematics education recommends first using finger counting, then concrete structured representations and, finally, mental representations of numbers to perform numerical operations. Taken together, these results reveal an important debate between neurocognitive and mathematics education research concerning the benefits and detriments of finger-based strategies for numerical development. In the present review, the rationale of both lines of evidence will be discussed.

Evidence from developmental as well as neuroscientific studies suggest that finger counting activity plays an important role in the acquisition of numerical skills in children. It has been claimed that this skill helps in building motor-based representations of number that continue to influence number processing well into adulthood, facilitating the emergence of number concepts from sensorimotor experience through a bottom-up process. The act of counting also involves the acquisition and use of a verbal number system of which number words are the basic building blocks. Using a Cognitive Developmental Robotics paradigm we present results of a modeling experiment on whether finger counting and the association of number words (or tags) to fingers, could serve to bootstrap the representation of number in a cognitive robot, enabling it to perform basic numerical operations such as addition. The cognitive architecture of the robot is based on artificial neural networks, which enable the r...

Does the existence of body representations undermine the explanatory role of the body? Or do certain types of representation depend so closely upon the body that their involvement in a cognitive task implicates the body itself? In the introduction of this special issue we explore lines of tension and complement that might hold between the notions of embodiment and body representations, which remain too often neglected or obscure. To do so, we distinguish two conceptions of embodiment that either put weight on the explanatory role of the body itself or body representations. We further analyse how and to what extent body representations can be said to be embodied. Finally, we give an overview of the full volume articulated around foundational issues (How should we define the notion of embodiment? To what extent and in what sense is embodiment compatible with representationalism? To what extent and in what sense are sensorimotor approaches similar to behaviourism?) and their applications in several cognitive domains (perception, concepts, selfhood, social cognition). Cognitive science of the last 20 years has seen an increasing focus on the body. Embodied cognitive science has become an industry. However, its unity is questionable (Clark 2008a, b), and its status as a competitor or ancestor to traditional cognitive science is unclear (for divergent views compare Thagard 2005; Varela et al. 1991). Still, a worry shared by many is that classical cognitive science puts too much of the mind's cognitive resources in the head. In its most radical form, this worry leads to a Rev.Phil.Psych.

In the field of convergence between research in autonomous machine construction and biological systems understanding it is usually argued that building robots for research on autonomy by replicating extant animals is a valuable strategy for engineering autonomous intelligent systems. In this paper we will address the very issue of animat construction, the rationale behind this, their current implementations and the value they are producing. It will be shown that current activity, as it is done today, is deeply flawed and ...

In this paper, we propose a formal definition of the perception as a behavioral dynamical attraction basin. The perception is built from the integration of the sensori-motor flow. Psychological considerations and robotic experiments on an embodied "intelligent" system are provided to show how this definition can satisfy both psychologist and robotician point of view.

The intelligence needed for autonomously controlling systems in high uncertainty conditions seems to be part of animal competences. An important part of the research in autonomous machine construction is focused in building robots by replicating extant animals. The question is if this is a valuable strategy for engineering autonomous intelligent systems. In this paper we will address the very issue of animat construction, the rationale behind this, their current implementations and the value they are producing for the ...

This paper is at the crossroad of Cognitive Psychology and AI Robotics. It reports a cross-disciplinary project concerned about implementing human heuristics within autonomous mobile robots. In the following, we address the problem of relying on human-based heuristics to endow a group of mobile robots with the ability to solve problems such as target finding in a labyrinth. Such heuristics may provide an efficient way to explore the environment and to decompose a complex problem into subtasks for which specific heuristics are efficient. We first present a set of experiments conducted with group of humans looking for a target with limited sensing capabilities solving. Then we describe the heuristics extracted from the observation and analysis of their behavior. Finally we implemented these heuristics within khepera-like autonomous mobile robots facing the same tasks. We show that the control architecture can be experimentally validated to some extent thanks to this approach.

Systems intended to operate in dynamic, complex environmentswithout intervention from their designers or significant amounts of domaindependent information provided at design time -must be equipped with a sufficient level of existential autonomy. This feature of naturally intelligent systems has largely been missing from cognitive architectures created to date, due in part to the fact that high levels of existential autonomy require systems to program themselves; good principles for self-programming have remained elusive. Achieving this with the major programming methodologies in use today is not likely, as these are without exception designed to be used by the human mind: Producing self-programming systems that can grow from first principles using these therefore requires first solving the AI problem itself -the very problem we are trying to solve. Advances in existential autonomy call for a new programming paradigm, with self-programming squarely at its center. The principles of such a paradigm are likely to be fundamentally different from prevailing approaches; among the desired features for a programming language designed for automatic self-programming are (a) support for autonomous knowledge acquisition, (b) real-time and any-time operation, (c) reflectivity, and (d) massive parallelization. With these and other requirements guiding our work, we have created a programming paradigm and language called Replicode. Here we discuss the reasoning behind our approach and the main motivations and features that set this work from apart from prior approaches.

This paper is at the crossroad of Cognitive Psychology and AI Robotics. It reports a cross-disciplinary project concerned about implementing human heuristics within autonomous mobile robots. In the following, we address the problem of relying on human-based heuristics to endow a group of mobile robots with the ability to solve problems such as target finding in a labyrinth. Such heuristics may provide an efficient way to explore the environment and to decompose a complex problem into subtasks for which specific heuristics are efficient. We first present a set of experiments conducted with group of humans looking for a target with limited sensing capabilities solving. Then we describe the heuristics extracted from the observation and analysis of their behavior. Finally we implemented these heuristics within khepera-like autonomous mobile robots facing the same tasks. We show that the control architecture can be experimentally validated to some extent thanks to this approach.

T he embodied cognition theory affirms that human intelligence is formed not only by the brain, but is also shaped by the body and the experiences acquired through it, such as manipulatives, gestures and movements . Research in developmental psychology has shown that embodied experiences help children in learning various cognitive skills by using limbs and senses to interact with the surrounding environment and other human beings 5 .

obtained his Ph.D. in computer science and artificial intelligence from the University of Queensland in Australia. He is currently a Research Fellow at the Queensland Brain Institute, where he has been analysing micro-electrode recording data from patients undergoing electrode implantation for deep brain stimulation for the treatment of brain disorders. One of his main aims is to understand This article is protected by copyright. All rights reserved. the computational principles which are supported by brain activity. New discoveries have the potential for multiple substantial benefits: primarily, revelations of brain function and dysfunction leading to new treatments for brain trauma and disease. Additionally, improved understanding of brain function could lead to the construction of novel, engineered information-processing and robotic systems to assist and augment human capabilities.

Thanks to recent technological advances and the increasing interest towards the Cognitive Developmental Robotics (CDR) paradigm, many popular platforms for scientific research have been designed in order to resemble the shape of the human body. The motivation behind this strongly humanoid design is the embodied cognition hypothesis, which affirms that all aspects of cognition are shaped by aspects of the body. Thus CDR is based on a synthetic approach that aims to provide new understanding on how human beings develop their higher cognitive functions. Following this paradigm we have developed an artificial model, based on artificial neural networks, to explore finger counting and the association of number words (or tags) to the fingers, as bootstrapping for the representation of numbers in the humanoid robot iCub. In this paper, we detail experiments of our model with the iCub robotic platform. Results of the number learning with proprioceptive data from the real platform are reported and compared with the ones obtained instead, with the simulated platform. These results support the thesis that learning the number words in sequence, along with finger configurations helps the building of the initial representation of number in the robot. Moreover, the comparison between the real and simulated iCub gives insights on the use of these platforms as a tool for CDR.