Guardado en:
Detalles Bibliográficos
Autores principales: Dang, Xuzhe, Edelkamp, Stefan
Formato: Preprint
Publicado: 2023
Materias:
Acceso en línea:https://arxiv.org/abs/2311.03485
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929690565935104
author Dang, Xuzhe
Edelkamp, Stefan
author_facet Dang, Xuzhe
Edelkamp, Stefan
contents This paper presents a novel method for learning reward functions for robotic motions by harnessing the power of a CLIP-based model. Traditional reward function design often hinges on manual feature engineering, which can struggle to generalize across an array of tasks. Our approach circumvents this challenge by capitalizing on CLIP's capability to process both state features and image inputs effectively. Given a pair of consecutive observations, our model excels in identifying the motion executed between them. We showcase results spanning various robotic activities, such as directing a gripper to a designated target and adjusting the position of a cube. Through experimental evaluations, we underline the proficiency of our method in precisely deducing motion and its promise to enhance reinforcement learning training in the realm of robotics.
format Preprint
id arxiv_https___arxiv_org_abs_2311_03485
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle CLIP-Motion: Learning Reward Functions for Robotic Actions Using Consecutive Observations
Dang, Xuzhe
Edelkamp, Stefan
Robotics
Artificial Intelligence
This paper presents a novel method for learning reward functions for robotic motions by harnessing the power of a CLIP-based model. Traditional reward function design often hinges on manual feature engineering, which can struggle to generalize across an array of tasks. Our approach circumvents this challenge by capitalizing on CLIP's capability to process both state features and image inputs effectively. Given a pair of consecutive observations, our model excels in identifying the motion executed between them. We showcase results spanning various robotic activities, such as directing a gripper to a designated target and adjusting the position of a cube. Through experimental evaluations, we underline the proficiency of our method in precisely deducing motion and its promise to enhance reinforcement learning training in the realm of robotics.
title CLIP-Motion: Learning Reward Functions for Robotic Actions Using Consecutive Observations
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2311.03485