从视觉感觉数据中控制人造代理是一项艰巨的任务。强化学习(RL)算法可以在这方面取得成功,但需要代理与环境之间进行大量相互作用。为了减轻该问题,无监督的RL建议采用自我监督的互动和学习,以更快地适应未来的任务。但是,目前的无监督策略是否可以改善概括能力,尤其是在视觉控制设置中。在这项工作中,我们为数据有效的视觉控制设计了有效的无监督RL策略。首先,我们表明,使用无监督的RL收集的数据预先训练的世界模型可以促进适应未来的任务。然后,我们与我们的混合计划者分析了一些设计选择,以有效地适应了代理的预训练组件,并在想象中学习和计划,并与我们的混合计划者一起使用,我们将其dub dyna-mpc进行了。通过结合一项大规模实证研究的发现,我们建立了一种方法,该方法强烈改善了无监督的RL基准测试的性能,需要20美元$ \ times $ $ $ $ $ \少于数据以符合监督方法的性能。该方法还表明了在现实词的RL基准测试上的稳健性能,暗示该方法概括为嘈杂的环境。
translated by 谷歌翻译
主动推断是一种特别是理解大脑的第一原理方法,通常是一种有情的药物,而自由能的单一命令。因此,它通过定义代理的生成模型并推断模型参数,动作和隐藏的状态信念,为对人工智能代理建模提供了一个计算帐户。但是,生成模型和隐藏状态空间结构的确切规范留给了实验者,其设计选择会影响代理的产生行为。最近,已经提出了深度学习方法,以从数据中学习隐藏的状态空间结构,从而从这项乏味的设计任务中减轻了实验者,但导致了一个纠缠的,不可解剖的状态空间。在本文中,我们假设这样一种学识渊博的,纠缠的状态空间并不一定会在自由能中产生最佳模型,并且在状态空间中执行不同的因素可以产生较低的模型复杂性。特别是,我们考虑了3D对象表示的问题,并专注于Shapenet数据集的不同实例。我们提出了一个分配对象形状,姿势和类别的模型,同时仍使用深层神经网络学习每个因素的表示形式。我们表明,当活跃代理在达到首选观察方面采用时,具有最佳分离属性的模型在采用时表现最好。
translated by 谷歌翻译
当研究不受限制的行为并允许小鼠离开笼子去驾驶复杂的迷宫时,小鼠在迷宫中表现出觅食行为,以寻求奖励,不时返回他们的家园,例如。喝。令人惊讶的是,当执行这样的``本垒打''时,老鼠不会遵循确切的反向路径,实际上,入口路径和家居路径几乎没有重叠。最近的工作提出了导航的层次主动推理模型,低级别模型对隐藏状态进行了推断,并提出了解释感官输入的姿势,而高级模型则可以推断出在位置之间移动,从而有效地构建环境地图。但是,使用此``MAP''进行计划,只允许代理找到它以前探索的轨迹,这与观察到的小鼠行为相去甚远。在本文中,我们探讨了通过使用低级生成模型来想象潜在的,但未发现的路径,探讨了将前路径纳入计划算法的方法。我们在网格世界环境中演示了概念证明,展示了代理如何使用从基于像素的观测值中学到的生成模型准确地预测地图中的新的,更短的路径。
translated by 谷歌翻译
自由能原理及其必然的积极推论构成了一种生物启发的理论,该理论假设生物学作用保留在一个受限制的世界首选状态中,即它们最小化自由能。根据这一原则,生物学家学习了世界的生成模型和未来的计划行动,该模型将使代理保持稳态状态,以满足其偏好。该框架使自己在计算机中实现,因为它理解了使其计算负担得起的重要方面,例如变异推断和摊销计划。在这项工作中,我们研究了深度学习的工具,以设计和实现基于主动推断的人造代理,对自由能原理进行深入学习的呈现,调查工作与机器学习和主动推理领域相关,以及讨论实施过程中涉及的设计选择。该手稿探究了积极推理框架的新观点,将其理论方面扎根于更务实的事务中,为活跃推理的新手提供了实用指南,并为深度学习从业人员的起点提供了研究,以调查自由能源原则的实施。
translated by 谷歌翻译
积极推论是一种统一的感知和行动理论,依赖于通过最小化自由能量来维持世界的内部模型。从行为的角度来看,有效推论代理商可以被视为自我证明的生命,以满足他们的乐观预测,即优选的结果或目标。相比之下,加固学习需要人工设计的奖励来完成任何期望的结果。尽管有效推理可以提供更自然的自我监控目标的控制,但其适用性因其在复杂环境中缩放方法的缺点而受到限制。在这项工作中,我们提出了对比主动推断的对比目标,这强烈降低了学习代理商的生成模式和规划未来行动的计算负担。我们的方法在基于图像的任务中的基于似的主动推断的情况下表现出显着优于基于似的主动推断,同时也是计算地更便宜,更容易训练。我们与能够获得人类设计奖励功能的加强学习代理,表明我们的方法与其表现完全符合。最后,我们还表明,在环境中的牵引力的情况下,对比方法显着更好地表现出明显更好,并且我们的方法能够将目标概括为背景中的变化。
translated by 谷歌翻译
Real-world robotic grasping can be done robustly if a complete 3D Point Cloud Data (PCD) of an object is available. However, in practice, PCDs are often incomplete when objects are viewed from few and sparse viewpoints before the grasping action, leading to the generation of wrong or inaccurate grasp poses. We propose a novel grasping strategy, named 3DSGrasp, that predicts the missing geometry from the partial PCD to produce reliable grasp poses. Our proposed PCD completion network is a Transformer-based encoder-decoder network with an Offset-Attention layer. Our network is inherently invariant to the object pose and point's permutation, which generates PCDs that are geometrically consistent and completed properly. Experiments on a wide range of partial PCD show that 3DSGrasp outperforms the best state-of-the-art method on PCD completion tasks and largely improves the grasping success rate in real-world scenarios. The code and dataset will be made available upon acceptance.
translated by 谷歌翻译
Artificial neural networks can learn complex, salient data features to achieve a given task. On the opposite end of the spectrum, mathematically grounded methods such as topological data analysis allow users to design analysis pipelines fully aware of data constraints and symmetries. We introduce a class of persistence-based neural network layers. Persistence-based layers allow the users to easily inject knowledge about symmetries (equivariance) respected by the data, are equipped with learnable weights, and can be composed with state-of-the-art neural architectures.
translated by 谷歌翻译
Quantifying motion in 3D is important for studying the behavior of humans and other animals, but manual pose annotations are expensive and time-consuming to obtain. Self-supervised keypoint discovery is a promising strategy for estimating 3D poses without annotations. However, current keypoint discovery approaches commonly process single 2D views and do not operate in the 3D space. We propose a new method to perform self-supervised keypoint discovery in 3D from multi-view videos of behaving agents, without any keypoint or bounding box supervision in 2D or 3D. Our method uses an encoder-decoder architecture with a 3D volumetric heatmap, trained to reconstruct spatiotemporal differences across multiple views, in addition to joint length constraints on a learned 3D skeleton of the subject. In this way, we discover keypoints without requiring manual supervision in videos of humans and rats, demonstrating the potential of 3D keypoint discovery for studying behavior.
translated by 谷歌翻译
Artificial intelligence is set to be deployed in operating rooms to improve surgical care. This early-stage clinical evaluation shows the feasibility of concurrently attaining real-time, high-quality predictions from several deep neural networks for endoscopic video analysis deployed for assistance during three laparoscopic cholecystectomies.
translated by 谷歌翻译
AI-based code generators are an emerging solution for automatically writing programs starting from descriptions in natural language, by using deep neural networks (Neural Machine Translation, NMT). In particular, code generators have been used for ethical hacking and offensive security testing by generating proof-of-concept attacks. Unfortunately, the evaluation of code generators still faces several issues. The current practice uses automatic metrics, which compute the textual similarity of generated code with ground-truth references. However, it is not clear what metric to use, and which metric is most suitable for specific contexts. This practical experience report analyzes a large set of output similarity metrics on offensive code generators. We apply the metrics on two state-of-the-art NMT models using two datasets containing offensive assembly and Python code with their descriptions in the English language. We compare the estimates from the automatic metrics with human evaluation and provide practical insights into their strengths and limitations.
translated by 谷歌翻译