Conflictual cues and unexpected changes in human real-case scenarios may be detrimental to the execution of tasks by artificial agents, thus affecting their performance. Meta-learning applied to reinforcement learning may enhance the design of control algorithms, where an outer learning system progressively adjusts the operation of an inner learning system, leading to practical benefits for the learning schema. Here, we developed a brain-inspired meta-learning framework for inhibition cognitive control that i) exploits the meta-learning principles in the neuromodulation theory proposed by Doya, ii) relies on a well-established neural architecture that contains distributed learning systems in the human brain, and iii) proposes optimization rules of meta-learning hyperparameters that mimic the dynamics of the major neurotransmitters in the brain. We tested an artificial agent in inhibiting the action command in two well-known tasks described in the literature: NoGo and Stop-Signal Paradigms. After a short learning phase, the artificial agent learned to react to the hold signal, and hence to successfully inhibit the motor command in both tasks, via the continuous adjustment of the learning hyperparameters. We found a significant increase in global accuracy, right inhibition, and a reduction in the latency time required to cancel the action process, i.e., the Stop-signal reaction time. We also performed a sensitivity analysis to evaluate the behavioral effects of the meta-parameters, focusing on the serotoninergic modulation of the dopamine release. We demonstrated that brain-inspired principles can be integrated into artificial agents to achieve more flexible behavior when conflictual inhibitory signals are present in the environment.

Predictive processing has become an influential framework in cognitive sciences. This framework turns the traditional view of perception upside down, claiming that the main flow of information processing is realized in a top-down, hierarchical manner. Furthermore, it aims at unifying perception, cognition, and action as a single inferential process. However, in the related literature, the predictive processing framework and its associated schemes, such as predictive coding, active inference, perceptual inference, and free-energy principle, tend to be used interchangeably. In the field of cognitive robotics, there is no clear-cut distinction on which schemes have been implemented and under which assumptions. In this letter, working definitions are set with the main aim of analyzing the state of the art in cognitive robotics research working under the predictive processing framework as well as some related nonrobotic models. The analysis suggests that, first, research in both cognitive robotics implementations and nonrobotic models needs to be extended to the study of how multiple exteroceptive modalities can be integrated into prediction error minimization schemes. Second, a relevant distinction found here is that cognitive robotics implementations tend to emphasize the learning of a generative model, while in nonrobotics models, it is almost absent. Third, despite the relevance for active inference, few cognitive robotics implementations examine the issues around control and whether it should result from the substitution of inverse models with proprioceptive predictions. Finally, limited attention has been placed on precision weighting and the tracking of prediction error dynamics. These mechanisms should help to explore more complex behaviors and tasks in cognitive robotics research under the predictive processing framework.

Multimodal information can significantly increase the perceptual capabilities of robotic agents, at the cost of a more complex sensory processing. This complexity can be reduced by employing machine learning techniques, provided that there is enough meaningful data to train on. This paper reports on creating novel datasets constructed by employing the iCub robot equipped with an additional depth sensor and color camera. We used the robot to acquire color and depth information for 210 objects in different acquisition scenarios. At the end, the results were large scale datasets that can be used for robot and computer vision applications: multisensory object representation, action recognition, rotation and distance invariant object recognition.

Hand-eye coordination is an important motor skill acquired in infancy which precedes pointing behavior. Pointing facilitates social interactions by directing attention of engaged participants. It is thus essential for the natural flow of human-robot interaction. Here, we attempt to explain how pointing emerges from sensorimotor learning of hand-eye coordination in a humanoid robot. During a body babbling phase with a random walk strategy, a robot learned mappings of joints for different arm postures. Arm joint configurations were used to train biologically inspired models consisting of SOMs. We show that such a model implemented on a robotic platform accounts for pointing behavior while humans present objects out of reach of the robot's hand.

We propose a computational model based on inverse-forward model pairs for the simulation and execution of actions. The models are implemented on a humanoid robot and are used to control reaching actions with the arms. In the experimental setup a tool has been attached to the left arm of the robot extending its covered action space. The preliminary investigations carried out aim at studying how the use of tools modifies the body scheme of the robot. The system performs action simulations before the actual executions. For each of the arms, predicted end-effector positions are compared with the desired one and the internal pair presenting the lowest error is selected for action execution. This allows the robot to decide on performing an action either with its hand alone or with the one with the attached tool.

Motor Babbling has been identified as a self-exploring behaviour adopted by infants and is fundamental for the development of more complex behaviours, self-awareness and social interaction skills. Here, we adopt this paradigm for the learning strategies of a humanoid robot that maps its random arm movements with its head movements, determined by the perception of its own body. Finally, we analyse three random movement strategies and experimentally test on a humanoid robot how they affect the learning speed.

Understanding and efficiently representing skills is one of the most important problems in a general Programming by Demonstration (PbD) paradigm. We present Growing Hierarchical Dynamic Bayesian Networks (GHDBN), an adaptive variant of the general DBN model able to learn and to represent complex skills. The structure of the model, in terms of number of states and possible transitions between them, is not needed to be known a priori. Learning in the model is performed incrementally and in an unsupervised manner.

Imitation learning has been recognized as a promising technique to teach robots advanced skills. It is based on the idea that robots could learn new behaviors by observing and imitating the behaviors of other skilled actors. We propose an adaptive probabilistic graphical model which copes with three core issues of any imitative behavior: observation, representation and reproduction of skills. Our model, Growing Hierarchical Dynamic Bayesian Network (GHDBN), is hierarchical (i.e. able to characterize structured behaviors at different levels of abstraction), and growing (i.e. skills are learned or updated incrementally - and at each level of abstraction - every time a new observation sequence is available). A GHDBN, once trained, is able to recognize skills being observed and to reproduce them by exploiting the generative power of the model. The system has been successfully tested in simulation, and initial tests have been conducted on a NAO humanoid robot platform.