arXiv 1904.074822019-04cs.LGcs.AIcs.CVstat.ML

基于多级抽象的面向对象动态学习

Object-Oriented Dynamics Learning through Multi-Level Abstraction

Institute for Interdisciplinary Information Sciences, Tsinghua University（清华大学交叉信息研究院）；Department of Computer Science and Technology, Tsinghua University（清华大学计算机科学与技术系）

Guangxiang Zhu, Jianhao Wang, Zhizhou Ren, Zichuan Lin, Chongjie Zhang

备注：: Accepted to the Thirthy-Fourth AAAI Conference On Artificial Intelligence (AAAI), 2020

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

摘要：

基于对象的方法学习动作条件动力学已被证明是有希望的推广和解释。然而，现有的方法遭受的结构限制和优化困难的共同环境中的多个动态对象。在本文中，我们提出了一种新的自监督学习框架，称为多级抽象面向对象的预测器（MAOP），它采用了三个层次的学习架构，使有效的基于对象的动态学习从原始的视觉观察。我们还为MAOP设计了一个时空关系推理机制，以支持实例级动态学习和处理部分可观测性。我们的研究结果表明，MAOP显着优于以前的方法，在样本效率和泛化的新环境的学习环境模型。我们还证明了学习的动态模型可以在看不见的环境中进行有效的规划，与真实的环境模型相当。此外，MAOP学习语义和视觉上可解释的分离表示。

英文摘要：

Object-based approaches for learning action-conditioned dynamics has demonstrated promise for generalization and interpretability. However, existing approaches suffer from structural limitations and optimization difficulties for common environments with multiple dynamic objects. In this paper, we present a novel self-supervised learning framework, called Multi-level Abstraction Object-oriented Predictor (MAOP), which employs a three-level learning architecture that enables efficient object-based dynamics learning from raw visual observations. We also design a spatial-temporal relational reasoning mechanism for MAOP to support instance-level dynamics learning and handle partial observability. Our results show that MAOP significantly outperforms previous methods in terms of sample efficiency and generalization over novel environments for learning environment models. We also demonstrate that learned dynamics models enable efficient planning in unseen environments, comparable to true environment models. In addition, MAOP learns semantically and visually interpretable disentangled representations.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

摘要：

英文摘要：

相关论文