Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing sensor data, e.g., laser sensor data, using neural networks. One of the methods includes obtaining a temporal sequence of multiple three-dimensional point clouds generated from sensor readings of an environment collected by one or more sensors within a given time period, each three-dimensional point cloud comprising a respective plurality of points in a first coordinate system; processing, using a feature extraction neural network, an input that comprises data derived from the temporal sequence of multiple three-dimensional point clouds to generate a feature embedding; receiving a query that specifies one time point within the given time period; and generating, from the feature embedding and conditioned on the query, one or more outputs that characterize one or more objects in the environment at the time point specified in the received query.

Abstract: In various embodiments, an unsupervised training application executes a neural network on a first point cloud to generate keys and values. The unsupervised training application generates output vectors based on a first query set, the keys, and the values and then computes spatial features based on the output vectors. The unsupervised training application computes quantized context features based on the output vectors and a first set of codes representing a first set of 3D geometry blocks. The unsupervised training application modifies the first neural network based on a likelihood of reconstructing the first point cloud, the quantized context features, and the spatial features to generate an updated neural network. A trained machine learning model includes the updated neural network, a second query set, and a second set of codes representing a second set of 3D geometry blocks and maps a point cloud to a representation of 3D geometry instances.

Abstract: In various embodiments, an unsupervised training application trains a machine learning model to generate representations of point clouds. The unsupervised training application executes a neural network on a first point cloud representing a first three-dimensional (3D) scene to generate segmentations. Based on the segmentations, the unsupervised training application computes spatial features. The unsupervised training application computes quantized context features based on the segmentations and a first set of codes representing a first set of 3D geometry blocks. The unsupervised training application modifies the neural network based on a likelihood of reconstructing the first point cloud, the quantized context features, and the spatial features to generate an updated neural network.

Abstract: In various embodiments, an unsupervised training application executes a neural network on a first point cloud to generate keys and values. The unsupervised training application generates output vectors based on a first query set, the keys, and the values and then computes spatial features based on the output vectors. The unsupervised training application computes quantized context features based on the output vectors and a first set of codes representing a first set of 3D geometry blocks. The unsupervised training application modifies the first neural network based on a likelihood of reconstructing the first point cloud, the quantized context features, and the spatial features to generate an updated neural network. A trained machine learning model includes the updated neural network, a second query set, and a second set of codes representing a second set of 3D geometry blocks and maps a point cloud to a representation of 3D geometry instances.