Search

With the development of IoT technology, more and more IoT devices are connected to the network. Due to the hardware constraints of IoT devices themselves, it is difficult for developers to embed security software into them. Therefore, it is better to protect IoT devices at the traffic level. The effect of malicious traffic detection based on neural networks is promising. Still, the slow computation brings some difficulties to deploying AI-based detection systems on edge servers. Time Convolutional Network (TCN) is a high-speed neural network suitable for massively parallel computation. In this paper, we propose Multi-class S-TCN, an improved network supporting multiple classifications based on TCN for the practical needs of IoT scenarios. Besides, we implement a complete IoT traffic security detection procedure based on deep packet inspection and protocol analysis. The proposed Multi-class S-TCN significantly improves the detection speed without degrading the detection effect. Experiments show that this work has better detection performance and faster detection speed compared to existing approaches, proving the effectiveness of the proposed detection flow and Multi-class S-TCN in IoT scenarios. [ABSTRACT FROM AUTHOR]

Federated learning (FL) plays an important role in the development of smart cities. With the evolution of big data and artificial intelligence, issues related to data privacy and protection have emerged, which can be solved by FL. In this paper, the current developments in FL and its applications in various fields are reviewed. With a comprehensive investigation, the latest research on the application of FL is discussed for various fields in smart cities. We explain the current developments in FL in fields, such as the Internet of Things (IoT), transportation, communications, finance, and medicine. First, we introduce the background, definition, and key technologies of FL. Then, we review key applications and the latest results. Finally, we discuss the future applications and research directions of FL in smart cities. [ABSTRACT FROM AUTHOR]

This paper explores the prospect of artificial intelligence (AI) supplementing or supplanting the human teacher in spiritual education using the example of Confucianism. The article compares the role and capabilities of two Confucian teachers: Confucius and an imagined robot teacher ("Digital Confucius"). It is argued in this article that Digital Confucius, if developed, would be useful as an "AI partner" in transmitting the knowledge and skills. As "AI mind", however, digital Confucius, would be weak in catering its teaching to suit the learner's profiles and contextual needs as well as serving as an ethical-spiritual guide and role-model to its students. Two major implications are discussed in this paper. First, the study highlights the potential complementary function of AI in the transfer of information and the irreplaceable role of the human teacher in moral education. Secondly, rather than making spirituality redundant, the limitation of AI in Confucian teaching speaks of the continuous relevance and centrality of spiritual development for human beings. [ABSTRACT FROM AUTHOR]

With the development of IoT technology, more and more IoT devices are connected to the network. Due to the hardware constraints of IoT devices themselves, it is difficult for developers to embed security software into them. Therefore, it is better to protect IoT devices at the traffic level. The effect of malicious traffic detection based on neural networks is promising. Still, the slow computation brings some difficulties to deploying AI-based detection systems on edge servers. Time Convolutional Network (TCN) is a high-speed neural network suitable for massively parallel computation. In this paper, we propose Multi-class S-TCN, an improved network supporting multiple classifications based on TCN for the practical needs of IoT scenarios. Besides, we implement a complete IoT traffic security detection procedure based on deep packet inspection and protocol analysis. The proposed Multi-class S-TCN significantly improves the detection speed without degrading the detection effect. Experiments show that this work has better detection performance and faster detection speed compared to existing approaches, proving the effectiveness of the proposed detection flow and Multi-class S-TCN in IoT scenarios. [ABSTRACT FROM AUTHOR]

Federated learning (FL) plays an important role in the development of smart cities. With the evolution of big data and artificial intelligence, issues related to data privacy and protection have emerged, which can be solved by FL. In this paper, the current developments in FL and its applications in various fields are reviewed. With a comprehensive investigation, the latest research on the application of FL is discussed for various fields in smart cities. We explain the current developments in FL in fields, such as the Internet of Things (IoT), transportation, communications, finance, and medicine. First, we introduce the background, definition, and key technologies of FL. Then, we review key applications and the latest results. Finally, we discuss the future applications and research directions of FL in smart cities. [ABSTRACT FROM AUTHOR]

Recent years have seen significant developments in Artificial Intelligence and we are now beginning to see some applications applied to sexual technologies. This has led to certain individuals and groups calling for a ban on the development of artificial sexual companions or "sexbots". This is largely due to fears that commercialising sex in this way could could re-enforce existing gender inequalities, and sexual objectification. We argue that one possible risk mitigation strategy could be to develop ethical safeguards into sexual robots. In this paper we explore the problem of implementing artificial ethical agents through a review of the relevant literature. We begin by exploring the motivation to the problem, and the science fiction that has inspired it. Following this, the practical, and philosophical implementation challenges are discussed before providing an overview of the current state of the art. The paper concludes by introducing the grand challenges of the field. [ABSTRACT FROM AUTHOR]

Clothing is one of the symbols of human civilisation. Clothing design is an art form that combines practicality and artistry. The Dunhuang clothes culture has a long history which represents ancient Chinese aesthetics. Artificial intelligence (AI) technology has been recently applied to multiple areas, which is also drawing increasing attention in fashion. However, little research has been done on the usage of AI for the creation of clothing, especially in traditional culture. It is challenging that the exploration of computer science and Dunhuang clothing design, which is a cross-history interaction between AI and Chinese classical culture. In this paper, we propose ClothGAN, which is an innovative framework for "designing" new patterns and styles of clothes based on generative adversarial network (GAN) and style transfer algorithm. Besides, we built the Dunhuang clothes dataset and conducted experiments to generate new patterns and styles of clothes with Dunhuang elements. We evaluated these clothing works generated from different models by computing inception score (IS), human prefer score (HPS) and generated score (IS and HPS). The results show that our framework outperformed others in these designing works. [ABSTRACT FROM AUTHOR]

This report summarises the papers, presentations and discussion of the Artificial Intelligence and Simulation of Behaviour special workshop re-evaluating the Engineering and Physical Science and Arts and Humanities Research Councils (EPSRC and AHRC) 2010 Principles of Robotics. We describe the call for papers re-examining the workshop, summarise the papers and discussions that took place, and the voting that lead to our workshop adopting a series of proposals for amending the original Principles. The workshop discussed and voted on 14 specific “amendments, additions, or reflections” on the Principles. Of these, 9 out 14 were adopted by majority vote, 6 receiving strong support (67% or more in favour), 1 majority support (53%), with several of the remaining receiving mixed support of between 33% and 47%. An important and unanimous conclusion of the workshop was that “the Principles should be amended through a thorough and inclusive process”. Adopted proposals also highlighted the need to “focus on the protection of humanity” from possible future risks created by AI and robotics, and to take into account how society is changing and adapting to technological advances. [ABSTRACT FROM AUTHOR]

The EPSRC's Principles of Robotics advises the implementation of transparency in robotic systems, however research related to AI transparency is in its infancy. This paper introduces the reader of the importance of having transparent inspection of intelligent agents and provides guidance for good practice when developing such agents. By considering and expanding upon other prominent definitions found in literature, we provide a robust definition of transparency as a mechanism to expose the decision-making of a robot. The paper continues by addressing potential design decisions developers need to consider when designing and developing transparent systems. Finally, we describe our new interactive intelligence editor, designed to visualise, develop and debug real-time intelligence. [ABSTRACT FROM AUTHOR]

This paper presents a procedure which composes music pieces through handling four layers in music, namely pitches, rhythms, dynamics, and timber. As an innovative feature, the procedure uses the combination of a genetic algorithm with a synergetic variable neighbourhood search. Uniform and one-point crossover operators as well as two mutation operators conduct the search in the employed genetic algorithm. The key point with these four operators is that the uniform crossover operator and the first mutation operator are indiscriminate, in the sense of using no knowledge of music theory, whereas the employed one-point crossover operator and the second mutation operator are musically informed. Music theory is used for finding the suitability of its generated pieces. The method starts with generating an initial sequence of pitches with a musically informed module and then calculates the suitability of the pitch sequence through the embedded rules. The employed genetic algorithm applies the variable neighbourhood search method to its generated offspring genomes for increasing their quality. Pieces can be composed in major, minor, and harmonic minor scales based on the user's request. As well as composing the main notes, the procedure generates up to three chord notes associated with each main note and plays the result in a novel multithreading environment through running four threads concurrently. [ABSTRACT FROM AUTHOR]

The thought-provoking EPSRC Principles of Robotics stem largely from the reflection on the extent to which robots can affect our lives. These comments highlight the fact that, while the principles may address to a good extent the present technological challenges, they appear to be less immediately suited for future technological and conceptual dares. The first part of the paper is dedicated to the search of the definition of what a robot is. Such a definition should offer the basic conceptual platform on which a normative endeavour, aiming to regulate robots in society, should be based. Concluding that the Principles offer no clear yet flexible insight into such a (meta-) definition, which would allow one to take into account the parameters of informed technological imagination and of envisaged social transformation, the second half of the paper highlights a number of regulatory points of tension. Such tensions, it is argued, stem largely from the absence of an appropriate conceptual platform, influencing negatively the extent to which the principles can be effective in guiding social, ethical, legal and scientific conduct. [ABSTRACT FROM PUBLISHER]

In two separate papers, Artificial Intelligence (AI)/Robotics researcher Guy Hoffman takes as a starting point that actors have been in the business of reverse engineering human behaviour for centuries. In this paper, I follow the similar trajectories of AI and acting theory (AT), looking at three primary questions, in the hope of framing a response to Hoffman's papers: (1) How are the problems of training a human to simulate a fictional human both similar to and different from training a machine to simulate a human? (2) How are the larger questions of AI design and architecture similar to the larger questions that still remain within the area of AT? (3) Is there anything in the work of AI design that might advance the work of acting theorists and practitioners? The paper explores the use of “swarm intelligence” in recent models of both AT and AI, and considers the issues of embodied cognition, and the kinds of intelligence that enhances or inhibits imaginative immersion for the actor, and concludes with a consideration of the ontological questions raised by the trend towards intersubjective, dynamic systems of generative thought in both AT and AI. [ABSTRACT FROM AUTHOR]

The random subspace method (RSM) has proved its excellence in numbers of pattern recognition tasks. However, the standard RSM is limited owing to the randomness in its feature selection procedure that is likely to lead to feature subset having poor class separability. In this paper, a proposal for a margin-based criterion has been presented for the evaluation of the true significance of the features, together with the true classification ability of base classifiers, so that both the training phase and integration phase of standard RSM could be enhanced. In the training phase, the random feature selection procedure is enhanced using a weighted random feature selection procedure, in order to improve the classification ability of the base classifier. In the integration phase, the simple majority voting strategy is enhanced using a weighted majority voting strategy for the purpose of assigning the base classifiers with poor classification ability to the lower voting weights. Experimental results on 30 benchmark datasets, together with 6 high-dimensional datasets prove that the recommended approach is capable of better providing classification ability to the usual classification task, in addition to the high-dimensional classification task. [ABSTRACT FROM AUTHOR]

The features related to the real world data may be redundant and erroneous in nature. The vital role of feature selection (FS) in handling such type of features cannot be ignored in the area of computational learning. The two most commonly used objectives for FS are the maximisation of the accuracy and minimisation of the number of features. This paper presents an Elitism-based Multi-objective Differential Evolution algorithm for FS and the novelty lies in the searching process which uses Minkowski Score (MS) and simultaneously optimises three objectives. The MS is considered as the third objective to keep track of the feature subset which is capable enough to produce a good classification result even if the average accuracy is poor. Extreme Learning Machine because of its fast learning speed and high efficiency has been considered with this multi-objective approach as a classifier for FS. Twenty-one benchmark datasets have been considered for performance evaluation. Moreover, the selected feature subsets are tested using 10-fold cross-validation. A comparative analysis of the proposed approach with two classical models, three single objective algorithms, and four multi-objective algorithms has been carried out to test the efficacy of the model. [ABSTRACT FROM AUTHOR]

Advanced autonomous artificial systems will need incremental learning and adaptive abilities similar to those seen in humans. Knowledge from biology, psychology and neuroscience is now inspiring new approaches for systems that have sensory-motor capabilities and operate in complex environments. Eye/hand coordination is an important cross-modal cognitive function, and is also typical of many of the other coordinations that must be involved in the control and operation of embodied intelligent systems. This paper examines a biologically inspired approach for incrementally constructing compact mapping networks for eye/hand coordination. We present a simplified node-decoupled extended Kalman filter for radial basis function networks, and compare this with other learning algorithms. An experimental system consisting of a robot arm and a pan-and-tilt head with a colour camera is used to produce results and test the algorithms in this paper. We also present three approaches for adapting to structural changes during eye/hand coordination tasks, and the robustness of the algorithms under noise are investigated. The learning and adaptation approaches in this paper have similarities with current ideas about neural growth in the brains of humans and animals during tool-use, and infants during early cognitive development. [ABSTRACT FROM AUTHOR]

Weight-perturbation (WP) algorithms for supervised and/or reinforcement learning offer improved biological plausibility over backpropagation because of their reduced circuitry requirements for realization in neural hardware. This paper explores the hypothesis that biological synaptic noise might serve as the substrate by which weight perturbation is implemented. We explore the basic synaptic noise hypothesis (BSNH), which embodies the weakest assumptions about the underlying neural circuitry required to implement WP algorithms. This paper identifies relevant biological constraints consistent with the BSNH, taxonomizes existing WP algorithms with regard to consistency with those constraints, and proposes a new WP algorithm that is fully consistent with the constraints. By comparing the learning effectiveness of these algorithms via simulation studies, it is found that all of the algorithms can support traditional neural network learning tasks and have similar generalization characteristics, although the results suggest a trade-off between learning efficiency and biological accuracy. [ABSTRACT FROM AUTHOR]

In this paper, we discuss the limitations of current evolutionary robotics models and we propose a new framework that might solve some of these problems and lead to art open-ended evolutionary process in hardware. More specifically, the paper describes a novel approach where the usual concepts of population, generations and fitness tire made implicit in the system. Individuals co-evolve embedded in their environment. Exploiting the sell-assembling capabilities of the (simulated) robots, the genotype of a successful individual can spread in the population. In this way, interesting behaviours emerge spontaneously, resulting in chasing and evading other individuals, collective obstacle avoidance and co-ordinated motion of self-assembled structures. [ABSTRACT FROM AUTHOR]

In autonomous agent systems, internal memory can be an important element to overcome the limitations of purely reactive agent behaviour. This paper presents an analysis of memory requirements for T-maze tasks well known as the road sign problem. In these tasks, a robot agent should make a decision about turning left or right at the T-junction in the approach corridor, depending on a history of perceptions. The robot agent in simulation can sense the light intensity influenced by light lamps placed on the bank of the wall. We apply the evolutionary multiobjective optimization approach to finite state controllers with two objectives: behaviour performance and memory size. Then the internal memory is quantified by counting internal states needed for the T- maze tasks in noisy environments. In particular, we focused on the influence of noise on internal memory and behaviour performance, and it is shown that state machines with variable thresholds can improve the performance with a hysteresis effect to filter out noise. This paper also provides an analysis of noise effect on perceptions and its relevance on performance degradation in state machines. [ABSTRACT FROM AUTHOR]

Developmental robotics is an emerging field located at the intersection of robotics, cognitive science and developmental sciences. This paper elucidates the main reasons and key motivations behind the convergence of fields with seemingly disparate interests, and shows why developmental robotics might prove to be beneficial for all fields involved. The methodology advocated is synthetic and two-pronged: on the one hand, it employs robots to instantiate models originating from developmental sciences; on the other hand, it aims to develop better robotic systems by exploiting insights gained from studies on ontogenetic development. This paper gives a survey of the relevant research issues and points to some future research directions. [ABSTRACT FROM AUTHOR]

In this paper we focus on how instructions for actions can be modelled in a self-organizing memory. Our approach draws from the concepts of regional distributed modularity and self-organization. We describe a self-organizing model that clusters action representations into different locations dependent on the body part they are related to. In the first case study we consider semantic representations of action verb meaning and then extend this concept significantly in a second case study by using actual sensor readings from our MIRA robot. Furthermore, we outline a modular model for a self-organizing robot action control system using language for instruction. Our approach for robot control using language incorporates some evidence related to the architectural and processing characteristics of the brain (Wermter et al. 2001b). This paper focuses on the neurocognitive clustering of actions and regional modularity for language areas in the brain. In particular, we describe a self-organizing network that realizes action clustering (Pulvermüller 2003). [ABSTRACT FROM AUTHOR]

The purpose of this paper is to propose a compound sine function neural network (NN) with continuous learning algorithm for the velocity and orientation angle tracking control of a mobile robot. Herein, two NN controllers embedded in the closed-loop control system are capable of on-line continuous learning and do not require any knowledge of the dynamics model. The neuron function of the hidden layer in the three-layer feed-forward network structure is on the basis of combining a sine function with a unipolar sigmoid function. In the NN algorithm, the weight values are only adjusted between the nodes in hidden layer and the output nodes, while the weight values between the input layer and the hidden layer are one, that is, constant, without the weight adjustment. The developed NN controllers have simple algorithm and fast learning convergence. Therefore, the proposed NN controllers can be suitable for the real-time tracking control of the mobile robots. The simulation results show that the proposed NN controller has better control performance in the tracking control of the mobile robot. The compound sine function NN provides a new way to solve tracking control problems for a mobile robot. [ABSTRACT FROM PUBLISHER]

This paper focuses on adaptive motor control in the kinematic domain. Several motor-learning strategies from the literature are adopted to kinematic problems: ‘feedback-error learning’, ‘distal supervised learning’, and ‘direct inverse modelling’ (DIM). One of these learning strategies, DIM, is significantly enhanced by combining it with abstract recurrent neural networks. Moreover, a newly developed learning strategy (‘learning by averaging’) is presented in detail. The performance of these learning strategies is compared with different learning tasks on two simulated robot setups (a robot-camera-head and a planar arm). The results indicate a general superiority of DIM if combined with abstract recurrent neural networks. Learning by averaging shows consistent success if the motor task is constrained by special requirements. [ABSTRACT FROM AUTHOR]

In this paper, we present an artificial vision system that is trained with a genetic algorithm for categorising five different kinds of images (letters) of different sizes. The system, which has a limited field of view, can move its eye so as to explore the images visually. The analysis of the system at the end of the training process indicates that correct categorisation is achieved by (1) exploiting sensory-motor coordination so as to experience stimuli that facilitate discrimination, and (2) integrating perceptual and/or motor information over time through a process of accumulation of partially conflicting evidence. We discuss our results with respect to the possible different strategies for categorisation and to the possible roles that action can play in perception. [ABSTRACT FROM AUTHOR]

While classic artificial intelligence systems still struggle to incorporate commonsense knowledge properly, situated and embodied artificial intelligence (SEAI) aims to build animats that acquire a common-sense understanding of the world via interactions between simulated brains, bodies and environments. Neuroscientists believe that much of this common sense involves predictive models for physical activities, but the transfer of sensorimotor skill knowledge to cognition is non-trivial, indicating that SEAI may meet a daunting challenge of its own. This paper considers the neurological bases for implicit procedural and explicit declarative common sense, and the possibilities for its transfer from the former to the latter. This helps assess the prospects for SEAI eventually to surpass GOFAI (good old-fashioned AI) in the quest for generally intelligent systems. [ABSTRACT FROM AUTHOR]

In this paper, we present the results of an experiment in which a collection of simulated robots that have been evolved for the ability to solve a collective navigation problem develop a communication system that allows them to co-operate better. The analysis of the results obtained indicates how evolving robots develop a non-trivial communication system and exploit different communication modalities. The results also indicate how the possibility of co-adapting the robots' individual and social/communicative behaviour plays a key role in the development of progressively more complex and effective individuals. [ABSTRACT FROM AUTHOR]

Highly recurrent neural networks can learn reverberating circuits called Cell Assemblies (CAs). These networks can be used to categorize input, and this paper explores the ability of CAs to learn hierarchical categories. A simulator, based on spiking fatiguing leaky integrators, is presented with instances of base categories. Learning is done using a compensatory Hebbian learning rule. The model takes advantage of overlapping CAs where neurons may participate in more than one CA. Using the unsupervised compensatory learning rule, the networks learn a hierarchy of categories that correctly categorize 97% of the basic level presentations of the input in our test. It categorizes 100% of the super-categories correctly. A larger hierarchy is learned that correctly categorizes 100% of base categories, and 89% of super-categories. It is also shown how novel subcategories gain default information from their super-category. These simulations show that networks containing CAs can be used to learn hierarchical categories. The network then can successfully categorize novel inputs. [ABSTRACT FROM AUTHOR]

During the past two decades, embodiment has become an important concept in many areas of cognitive science, but so far there is no common understanding of what constitutes embodied cognition and what kind of body an artificial humanlike cognizer would require. Work in embodied artificial intelligence and robotics has addressed, to some degree, what kind of bodily implementation is necessary for embodied cognition, but crucial factors such as the role of social interaction and the ‘body-in-motion’ have still not received much attention. We argue that, in the human child, the interplay of social scaffolding and self-induced locomotion is fundamental to the development of joint attention and a ‘self’. Furthermore, we discuss the implications of the social dynamics of bodily experience for android science. We argue that keeping scientific and engineering perspectives apart, but also understanding their relation, is important for clarifying the objectives of android science.Parts of this paper have appeared in similar form in Lindblom and Ziemke (2005a, b) and Ziemke and Lindblom (2006). [ABSTRACT FROM AUTHOR]

What dynamics do simple recurrent networks (SRNs) develop to represent stack-like and queue-like memories? SRNs have been widely used as models in cognitive science. However, they are interesting in their own right as non-symbolic computing devices from the viewpoints of analogue computing and dynamical systems theory. In this paper, SRNs are trained on two prototypical formal languages with recursive structures that need stack-like or queue-like memories for processing, respectively. The evolved dynamics are analysed, then interpreted in terms of simple dynamical systems, and the different ease with which SRNs aquire them is related to the properties of these simple dynamical systems. Within the dynamical systems framework, it is concluded that the stack-like language is simpler than the queue-like language, without making use of arguments from symbolic computation theory. [ABSTRACT FROM AUTHOR]

Many researchers in embodied cognitive science and artificial intelligence, and evolutionary robotics in particular, emphasize the interaction or brain, body and environment as crucial to the emergence of intelligent. adaptive behaviour. Accordingly. the interaction between agent and environment. as well as the co-adaptation of artificial brains and bodies, has been the focus of much research in evolutionary robotics. Hence, there are plenty of studies of robotic agents/species adapting to a given environment. Many animals, on the other hand, in particular humans, to some extent can choose to adapt the environment to their own needs instead of adapting (onIy) themselves. That alternative has been studied relatively little in robot experiments. This paper, therefore, presents some simple initial simulation experiments, in a delayed response task setting, that illustrate how the evolution of environment adaptation can sent to provide cognitive scaffolding that reduces the requirements for individual agents. Furthermore, theoretical implications, open questions and future research directions for evolutionary robotics are discussed. [ABSTRACT FROM AUTHOR]

Microscopic analysis is a standard approach in the study of robot behaviour. Typically, the approach comprises the analysis of a single (or sometimes a few) robot-environment system(s) to reveal specific properties of robot behaviour. In contrast to microscopic analysis, macroscopic analysis focuses on averaged properties of systems. The advantage is that such a property is easier to generalize so that it can be established to what extent the property is universal. This paper investigates whether a macroscopic analysis can reveal a universal property of adaptive behaviour in a robot model of foraging behaviour. Our analysis reveals that the step lengths of the most successful robots are distributed according to a Lévy-flight distribution. From studies on a variety of natural species, it is known that such a distribution constitutes a universal property of foraging behaviour. Thereafter, we discuss an example of how macroscopic analysis can be applied to existing research in evolutionary robotics, and relate the macroscopic and microscopic analyses of foraging behaviour to the framework of scientific research described by Cohen (1995, Empirical Methods for Artificial Intelligence (Cambridge MA: MIT Press)). We conclude that macroscopic analysis may predict universal properties of adaptive behaviour and that it may complement microscopic analysis in the study of adaptive behaviour. [ABSTRACT FROM AUTHOR]

Berkeley et al. (1995, Connection Science, 7: 167–186) introduced a novel technique for analysing the hidden units of connectionist networks that had been trained using the backpropagation learning procedure. The literature concerning banding analysis is equivocal with respect to the kinds of processing units this technique can be used on. In this paper, it will be shown that, contrary to the claims in some published sources, banding analysis can be conducted on networks that use standard processing units that have a sigmoid activation function. The analytic process is then illustrated and the potential benefits of this kind of technique are discussed. [ABSTRACT FROM AUTHOR]

While humans forget gradually, highly distributed connectionist networks forget catastrophically: newly learned information often completely erases previously learned information. This is not just implausible cognitively, but disastrous practically. However, it is not easy in connectionist cognitive modelling to keep away from highly distributed neural networks, if only because of their ability to generalize. A realistic and effective system that solves the problem of catastrophic interference in sequential learning of ‘static’ (i.e. non-temporally ordered) patterns has been proposed recently (Robins 1995. Connection Science. 7: 123–146, 1996. Connection Science, 8: 259–275, Ans and Rousset 1997. CR Académie ties Sciences Paris. Life Sciences, 320: 989–997, French 1997, Connection Science. 9: 353–379, 1999, Trends in Cognitive Sciences, 3: 128–135, Ans and Rousset 2000, Connection Science, 12: 1–19). The basic principle is to learn new external patterns interleaved with internally generated ‘pseudopatterns’ (generated from random activation) that reflect the previously learned information. However, to be credible, this self-refreshing mechanism for static learning has to encompass our human ability to learn serially many temporal sequences of patterns without catastrophic forgetting. Temporal sequence learning is arguably more important than static pattern learning in the real world. In this paper, we develop a dual-network architecture in which self-generated pseudopatterns reflect (non-temporally) all the sequences of temporally ordered items previously learned. Using these pseudopatterns, several self-refreshing mechanisms that eliminate catastrophic forgetting in sequence learning are described and their efficiency is demonstrated through simulations. Finally, an experiment is presented that evidences a close similarity between human and simulated behaviour. [ABSTRACT FROM AUTHOR]

One of the most important abilities for an agent's cognitive development in a social environment is the ability to recognize and imitate actions of others. In this paper we describe a cognitive architecture for action recognition and imitation, and present experiments demonstrating its implementation in robots. Inspired by neuroscientific and psychological data, and adopting a 'simulation theory of mind' approach, the architecture uses the motor systems of the imitator in a dual role, both for generating actions, and for understanding actions when performed by others. It consists of a distributed system of inverse and forward models that uses prediction accuracy as a means to classify demonstrated actions. The architecture is also shown to be capable of learning new composite actions from demonstration. [ABSTRACT FROM AUTHOR]

This paper presents a constructive model by which a robot acquires the ability of joint attention with a human caregiver based on its embedded mechanisms of visual attention and learning with self-evaluation. The former is to look at a salient object in the robot's view, and the latter is to learn sensorimotor co-ordination when visual attention has succeeded. Since the success of visual attention does not always correspond to the success of joint attention, the robot has incorrect learning data for joint attention as well as correct data. However, the robot is expected statistically to lose incorrect data as outliers since such data do not have any correlation in the sensorimotor co-ordination while correct data have a correlation. The robot consequently acquires the ability of joint attention by finding the correlation in the sensorimotor co-ordination even if multiple objects are placed at random positions in an environment and a human caregiver does not provide any task evaluation to the robot. The experimental results show that the proposed model makes the robot reproduce the developmental process of infants' joint attention. Therefore, the proposed model could be one of the models to explain how infants develop the ability of joint attention. [ABSTRACT FROM AUTHOR]

Studies on emotion are currently receiving a lot of attention. The importance of emotion in the development and support of intelligent and social behaviour has been highlighted by studies in psychology and neurology. Hence, the recognition of affective states has also become a critical feature in robot social development, with robots assumed to take on a role as social companion. In this paper, we address the issue of endowing robots with the ability to learn incrementally to recognize the affective state of their human partner by interpreting their gestural cues. We propose a model that can self-organize postural features into affective categories, and use contextual feedback from the partner to drive the learning process. [ABSTRACT FROM AUTHOR]

This paper provides a summary of new results coming from developmental infancy research demonstrating preverbal infants' early competence in understanding and learning from the intentional actions of other agents. The reviewed studies (using violation-of-expectation and observational learning paradigms) provide converging evidence that by the end of the first year infants can interpret and draw systematic inferences about other agents' goal-directed actions, and can rely on such inferences in observational learning when imitating others' actions or emulating their goals. To account for these findings it is proposed that young infants possess a non-mentalistic action interpretational system, the 'teleological stance' (Gergely and Csibra 2003) that represents actions by relating three relevant aspects of reality (action, goal-state, and situational constraints) through the inferential 'principle of rational action', which assumes that: (a) the basic function of actions is to bring about future goal-states; and that (b) agents will always perform the most efficient means action available to them within the constraints of the given situation. The relevance of these research findings and the proposed theoretical model for how to realize the future goal of epigenetic robotics of building a 'socially relevant' humanoid robot is then discussed. [ABSTRACT FROM AUTHOR]

In this paper we address the problem of synthesizing mobile robots able to solve problems in which they cannot merely react to sensory input, but have to maintain an internal state as well. More precisely, we shall show how autonomous robots synthesized through an evolutionary process can solve problems that necessarily require an ability to integrate sensorimotor information over time. By presenting the result of a set of experiments in which evolving robots are asked to navigate and selflocalize in the environment, we shall show that successful results can be achieved by providing evolving individuals with neural controllers with neurons that: (a) vary their activity at different rates to detect regularities at different time-scales in the sensorimotor flow: and (b) use thresholded activation functions to detect events extending over time. [ABSTRACT FROM AUTHOR]

This paper introduces a novel networking strategy for RAM-based neurons which significantly improves the training and recognition performance of such networks whilst maintaining the generalization capabilities achieved in previous network configurations. A number of different architectures are introduced, each using the same underlying principles. Initially, features which are common to all architectures are described, illustrating the basis of the underlying paradigm. Three architectures are then introduced illustrating different techniques of employing the paradigm to meet differing performance specifications. The architectures are described in terms of the structure of the neurons they employ. Details of the various training and recognition algorithms employed by the architectures are supplied in order to present a complete description of the operation of this class of artificial neural network. [ABSTRACT FROM AUTHOR]

This paper presents a novel agent architecture, the VM-architecture, that chooses its actions via a behaviour instantiation mechanism inspired by Maes’s behaviour networks. The VM-architecture enables an agent, VMattie, built upon it, actively to gather information from humans using email in natural language, to compose announcements of next week’s seminars and to mail them each week to a list that she keeps updated. VMattie does all this without the supervision of a human. Implemented and tested in a UNIX environment, the system’s performance surpassed the authors’ fondest hopes. [ABSTRACT FROM AUTHOR]