Abstract: A method performed by one or more computers. The method comprises receiving a natural language query specifying requirements for a compound; processing the natural language query using a language policy to generate a plurality of representations of candidate compounds that each satisfy at least a subset of the requirements specified in the natural language query. Each representation specifies at least a chemical formula of the corresponding candidate compound. The method further comprises, for each representation in a subset of the representations, using a generative machine learning model conditioned on the representation to generate one or more candidate chemical structures, each candidate chemical structure comprising a respective spatial location for each of the atoms of the corresponding candidate compound; and selecting a chemical structure and corresponding compound from the plurality of candidate chemical structures.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for controlling an agent interacting with an environment. In one aspect, a method comprises: receiving one or more observations of an environment; receiving an input text sequence that describes a task to be performed by a robot in the environment; generating an encoded representation of the input text sequence in an embedding space; generating a corresponding encoded representation of each of the one or more observations in the embedding space; generating a sequence of input tokens that comprises the encoded representation of the input text sequence and the corresponding encoded representation of each observation; processing the sequence of input tokens using a language model neural network to generate an output text sequence that comprises high-level natural language instructions; and determining, from the high-level natural language instructions, one or more actions to be performed by the robot.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for controlling an agent interacting with an environment. In one aspect, a method comprises: receiving one or more observations of an environment; receiving an input text sequence that describes a task to be performed by a robot in the environment; generating an encoded representation of the input text sequence in an embedding space; generating a corresponding encoded representation of each of the one or more observations in the embedding space; generating a sequence of input tokens that comprises the encoded representation of the input text sequence and the corresponding encoded representation of each observation; processing the sequence of input tokens using a language model neural network to generate an output text sequence that comprises high-level natural language instructions; and determining, from the high-level natural language instructions, one or more actions to be performed by the robot.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system allows a user to move up and down within a scene relative to an up axis of the scene without the user being displaced horizontally. A scale is provided that shows the user the relative position of the current view in the vertical bounding box of the scene.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system provides allows a user to place a point on a model surface that is used as a focus point for tool operations. The focus point is a geometry sphere that shows orientation and relative position of the view in the scene.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system provides scale GUIs that are positioned in the view and show the view position in the scene.

Abstract: A static model is populated with graphics objects by identifying a first graphics object that is associated with the static model, creating a first plurality of graphics objects where each graphics object in the first plurality of graphics objects comprises an instance of the first graphics object, placing each graphics object in the first plurality of graphics objects into a respective first position, and simulating a motion path for each graphics object in the first plurality of graphics objects from their respective first position to a respective second position.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system provides a rewind function where view waypoints are stored as navigation occurs and the user can move the view back to one of the rewind points.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system allows a user to designate a point on a model and the system moves the view toward that point. A perspective scale is provided that shows the user the relative position of the current view from the starting point of the view to the surface point.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system allows a user to designate a point on a model in a scene and the point is moved to the center of the view.

Abstract: An animation processor comprises a processor, an instruction memory, a working memory, inputs and outputs, and an input for receiving an input animation sequence comprising a plurality of time-sequenced frames, storage for a style image pair, comprising an input style image and an output style image, wherein differences between the input style image and an output style image convey a style for drawings and other visual elements, logic for generating set of time-varying orientation fields, logic for rendering velocity fields for the sequences, and an output for outputting an output animation sequence that takes on the style specified by the style image pair, with the output animation sequence having temporal coherence.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system provides a set of GUI tracking menus for different navigation tasks where each navigation tool has action tools associated with the navigation task. The action tools are arranged in rings with the most used tools on an outside.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system provides safe navigation features including tool clustering, orientation indicators in the scene, slider indicating scene position, motion warning graphics and other features that help in navigating the scene.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system provides a zoom tool that allows a user to designate a target point in a scene and zoom relative to that target point.

Abstract: One embodiment of the present invention sets forth a technique for providing an end user with a multiscale three-dimensional (3D) navigation experience in design software application programs. An adaptive multiscale 3D navigation system allows an end user to transition between a planetary scale down to an individual building scale. The end user may navigate within the building, inspecting object details within the building. The size of the environment is sensed automatically, and the viewing and travel parameters are adjusted accordingly to provide the end user with a seamless navigation experience. A consistent navigation experience is supported at various scales, and real-time collision detection is provided. Scale computation for 3D scenes and collision detection may be based on a generated depth cubemap of the environment.

Abstract: One embodiment of the present invention sets forth a technique for providing an end user with a multiscale three-dimensional (3D) navigation experience in design software application programs. An adaptive multiscale 3D navigation system allows an end user to transition between a planetary scale down to an individual building scale. The end user may navigate within the building, inspecting object details within the building. The size of the environment is sensed automatically, and the viewing and travel parameters are adjusted accordingly to provide the end user with a seamless navigation experience. A consistent navigation experience is supported at various scales, and real-time collision detection is provided. Scale computation for 3D scenes and collision detection may be based on a generated depth cubemap of the environment.

Abstract: A static model is populated with graphics objects by identifying a first graphics object that is associated with the static model, creating a first plurality of graphics objects where each graphics object in the first plurality of graphics objects comprises an instance of the first graphics object, placing each graphics object in the first plurality of graphics objects into a respective first position, and simulating a motion path for each graphics object in the first plurality of graphics objects from their respective first position to a respective second position.

Abstract: A navigation system for navigating a three-dimensional (3D) scene that includes a model or object with which a user can interact. The system accommodates and helps both novice and advanced users. To do this, the system locks a cursor to a model point in a scene during panning and controls panning speed relative to a bounding box size of the scene.

Abstract: One embodiment of the present invention sets forth a technique for providing an end user with a multiscale three-dimensional (3D) navigation experience in design software application programs. An adaptive multiscale 3D navigation system allows an end user to transition between a planetary scale down to an individual building scale. The end user may navigate within the building, inspecting object details within the building. The size of the environment is sensed automatically, and the viewing and travel parameters are adjusted accordingly to provide the end user with a seamless navigation experience. A consistent navigation experience is supported at various scales, and real-time collision detection is provided. Scale computation for 3D scenes and collision detection may be based on a generated depth cubemap of the environment.

Abstract: One embodiment of the present invention sets forth a technique for providing an end user with a multiscale three-dimensional (3D) navigation experience in design software application programs. An adaptive multiscale 3D navigation system allows an end user to transition between a planetary scale down to an individual building scale. The end user may navigate within the building, inspecting object details within the building. The size of the environment is sensed automatically, and the viewing and travel parameters are adjusted accordingly to provide the end user with a seamless navigation experience. A consistent navigation experience is supported at various scales, and real-time collision detection is provided. Scale computation for 3D scenes and collision detection may be based on a generated depth cubemap of the environment.