Language modelling provides a step towards intelligent communication systems by harnessing large repositories of written human knowledge to better predict and understand the world. In this paper, we present an analysis of Transformer-based language model performance across a wide range of model scales -from models with tens of millions of parameters up to a 280 billion parameter model called Gopher. These models are evaluated on 152 diverse tasks, achieving state-of-the-art performance across the majority. Gains from scale are largest in areas such as reading comprehension, fact-checking, and the identification of toxic language, but logical and mathematical reasoning see less benefit. We provide a holistic analysis of the training dataset and model's behaviour, covering the intersection of model scale with bias and toxicity. Finally we discuss the application of language models to AI safety and the mitigation of downstream harms.
Abstract-This article combines programming by demonstration and adaptive control for teaching a robot to physically interact with a human in a collaborative task requiring sharing of a load by the two partners. Learning a task model allows the robot to anticipate the partner's intentions and adapt its motion according to perceived forces. As the human represents a highly complex contact environment, direct reproduction of the learned model may lead to sub-optimal results. To compensate for unmodelled uncertainties, in addition to learning we propose an adaptive control algorithm that tunes the impedance parameters, so as to ensure accurate reproduction. To facilitate the illustration of the concepts introduced in this paper and provide a systematic evaluation, we present experimental results obtained with simulation of a dyad of two planar 2-DOF robots.
Abstract-Motion imitation requires reproduction of a dynamical signature of a movement, i.e. a robot should be able to encode and reproduce a particular path together with a specific velocity and/or an acceleration profile. Furthermore, a human provides only few demonstrations, that cannot cover all possible contexts in which the robot will need to reproduce the motion autonomously. Therefore, the encoding should be able to efficiently generalize knowledge by generating similar motions in unseen context. This work follows a recent trend in Programming by Demonstration in which the dynamics of the motion is learned. We present an algorithm to estimate multivariate robot motions through a Mixture of Gaussians.The strengths of the proposed encoding are three-fold: i) it allows to generalize a motion to unseen context; ii) it provides fast on-line replanning of the motion in the face of spatio-temporal perturbations; iii) it may embed different types of dynamics, governed by different attractors.The generality of the method to estimate arbitrary nonlinear motion dynamics is demonstrated by accurately estimating a set of known non-linear dynamical systems. The platformindependency and real-time performance of the method are further validated to learn the non-linear motion dynamics of manipulation tasks with different robotic platforms. We provide an experimental comparison of our approach with an related state-of-the-art method.
Abstract-This paper presents the application of a statistical framework that allows to endow a humanoid robot with the ability to perform a collaborative manipulation task with a human operator. We investigate to what extent the dynamics of the motion and the haptic communication process that takes place during physical collaborative tasks can be encapsulated by the probabilistic model. This framework encodes the dataset in a Gaussian Mixture Model, which components represent the local correlations across the variables that characterize the task. A set of demonstrations is performed using a bilateral coupling teleoperation setup; then the statistical model is trained in a pure follower/leader role distribution mode between the human and robot alternatively. The task is reproduced using Gaussian Mixture Regression. We present the probabilistic model and the experimental results obtained on the humanoid platform HRP-2; preliminary results assess our theory on switching behavior modes in dyad collaborative tasks: when reproduced with users which were not instructed to behave in either a follower or a leader mode, the robot switched automatically between the learned leader and follower behaviors.
Abstract-We present a generic framework that allows learning non-linear dynamics of motion in manipulation tasks and generating dynamical laws for control of position and orientation. This work follows a recent trend in Programming by Demonstration in which the dynamics of an arm motion is learned: position and orientation control are learned as multivariate dynamical systems to preserve correlation within the signals. The strength of the method is three-fold: i) it extracts dynamical control laws from demonstrations, and subsequently provides concurrent smooth control of both position and orientation; ii) it allows to generalize a motion to unseen context; iii) it guarantees on-line adaptation of the motion in the face of spatial and temporal perturbations. The method is validated to control a four degree of freedom humanoid arm and an industrial six degree of freedom robotic arm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.