Dataset Information

Intrinsically Motivated Exploration of Learned Goal Spaces.

ABSTRACT: Finding algorithms that allow agents to discover a wide variety of skills efficiently and autonomously, remains a challenge of Artificial Intelligence. Intrinsically Motivated Goal Exploration Processes (IMGEPs) have been shown to enable real world robots to learn repertoires of policies producing a wide range of diverse effects. They work by enabling agents to autonomously sample goals that they then try to achieve. In practice, this strategy leads to an efficient exploration of complex environments with high-dimensional continuous actions. Until recently, it was necessary to provide the agents with an engineered goal space containing relevant features of the environment. In this article we show that the goal space can be learned using deep representation learning algorithms, effectively reducing the burden of designing goal spaces. Our results pave the way to autonomous learning agents that are able to autonomously build a representation of the world and use this representation to explore the world efficiently. We present experiments in two environments using population-based IMGEPs. The first experiments are performed on a simple, yet challenging, simulated environment. Then, another set of experiments tests the applicability of those principles on a real-world robotic setup, where a 6-joint robotic arm learns to manipulate a ball inside an arena, by choosing goals in a space learned from its past experience.

SUBMITTER: Laversanne-Finot A

PROVIDER: S-EPMC7835425 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Intrinsically Motivated Exploration of Learned Goal Spaces.

Laversanne-Finot Adrien A Péré Alexandre A Oudeyer Pierre-Yves PY

Frontiers in neurorobotics 20210112

Finding algorithms that allow agents to discover a wide variety of skills efficiently and autonomously, remains a challenge of Artificial Intelligence. Intrinsically Motivated Goal Exploration Processes (IMGEPs) have been shown to enable real world robots to learn repertoires of policies producing a wide range of diverse effects. They work by enabling agents to autonomously sample goals that they then try to achieve. In practice, this strategy leads to an efficient exploration of complex environ ...[more]

PMID: 33510630

Similar Datasets

Project description:Over the last 20 years, a significant part of the research in exploratory robotics partially switches from looking for the most efficient way of exploring an unknown environment to finding what could motivate a robot to autonomously explore it. Moreover, a growing literature focuses not only on the topological description of a space (dimensions, obstacles, usable paths, etc.) but rather on more semantic components, such as multimodal objects present in it. In the search of designing robots that behave autonomously by embedding life-long learning abilities, the inclusion of mechanisms of attention is of importance. Indeed, be it endogenous or exogenous, attention constitutes a form of intrinsic motivation for it can trigger motor command toward specific stimuli, thus leading to an exploration of the space. The Head Turning Modulation model presented in this paper is composed of two modules providing a robot with two different forms of intrinsic motivations leading to triggering head movements toward audiovisual sources appearing in unknown environments. First, the Dynamic Weighting module implements a motivation by the concept of Congruence, a concept defined as an adaptive form of semantic saliency specific for each explored environment. Then, the Multimodal Fusion and Inference module implements a motivation by the reduction of Uncertainty through a self-supervised online learning algorithm that can autonomously determine local consistencies. One of the novelty of the proposed model is to solely rely on semantic inputs (namely audio and visual labels the sources belong to), in opposition to the traditional analysis of the low-level characteristics of the perceived data. Another contribution is found in the way the exploration is exploited to actively learn the relationship between the visual and auditory modalities. Importantly, the robot-endowed with binocular vision, binaural audition and a rotating head-does not have access to prior information about the different environments it will explore. Consequently, it will have to learn in real-time what audiovisual objects are of "importance" in order to rotate its head toward them. Results presented in this paper have been obtained in simulated environments as well as with a real robot in realistic experimental conditions.

Project description:The mid- and high-level visual properties supporting object perception in the ventral visual pathway are poorly understood. In the absence of well-specified theory, many groups have adopted a data-driven approach in which they progressively interrogate neural units to establish each unit's selectivity. Such methods are challenging in that they require search through a wide space of feature models and stimuli using a limited number of samples. To more rapidly identify higher-level features underlying human cortical object perception, we implemented a novel functional magnetic resonance imaging method in which visual stimuli are selected in real-time based on BOLD responses to recently shown stimuli. This work was inspired by earlier primate physiology work, in which neural selectivity for mid-level features in IT was characterized using a simple parametric approach (Hung et al., 2012). To extend such work to human neuroimaging, we used natural and synthetic object stimuli embedded in feature spaces constructed on the basis of the complex visual properties of the objects themselves. During fMRI scanning, we employed a real-time search method to control continuous stimulus selection within each image space. This search was designed to maximize neural responses across a pre-determined 1 cm(3) brain region within ventral cortex. To assess the value of this method for understanding object encoding, we examined both the behavior of the method itself and the complex visual properties the method identified as reliably activating selected brain regions. We observed: (1) Regions selective for both holistic and component object features and for a variety of surface properties; (2) Object stimulus pairs near one another in feature space that produce responses at the opposite extremes of the measured activity range. Together, these results suggest that real-time fMRI methods may yield more widely informative measures of selectivity within the broad classes of visual features associated with cortical object representation.

Project description:The plasticity of the human nervous system allows us to acquire an open-ended repository of sensorimotor skills in adulthood, such as the mastery of tools, musical instruments or sports. How novel sensorimotor skills are learned from scratch is yet largely unknown. In particular, the so-called inverse mapping from goal states to motor states is underdetermined because a goal can often be achieved by many different movements (motor redundancy). How humans learn to resolve motor redundancy and by which principles they explore high-dimensional motor spaces has hardly been investigated. To study this question, we trained human participants in an unfamiliar and redundant visually-guided manual control task. We qualitatively compare the experimental results with simulation results from a population of artificial agents that learned the same task by Goal Babbling, which is an inverse-model learning approach for robotics. In Goal Babbling, goal-related feedback guides motor exploration and thereby enables robots to learn an inverse model directly from scratch, without having to learn a forward model first. In the human experiment, we tested whether different initial conditions (starting positions of the hand) influence the acquisition of motor synergies, which we identified by Principal Component Analysis in the motor space. The results show that the human participants' solutions are spatially biased towards the different starting positions in motor space and are marked by a gradual co-learning of synergies and task success, similar to the dynamics of motor learning by Goal Babbling. However, there are also differences between human learning and the Goal Babbling simulations, as humans tend to predominantly use Degrees of Freedom that do not have a large effect on the hand position, whereas in Goal Babbling, Degrees of Freedom with a large effect on hand position are used predominantly. We conclude that humans use goal-related feedback to constrain motor exploration and resolve motor redundancy when learning a new sensorimotor mapping, but in a manner that differs from the current implementation of Goal Babbling due to different constraints on motor exploration.

Dataset Information

Intrinsically Motivated Exploration of Learned Goal Spaces.

Publications

Intrinsically Motivated Exploration of Learned Goal Spaces.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets