Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating compressed representations of synthetic images. One of the methods is a method of generating a synthetic image using a generative neural network, and includes: generating, using the generative neural network, a plurality of coefficients that represent the synthetic image after the synthetic image has been encoded using a lossy compression algorithm; and decoding the synthetic image by applying the lossy compression algorithm to the plurality of coefficients.

Abstract: An improved computing system can use mobility data corresponding to a particular mobile device to identify one or more geographic locations to which the mobile device traveled within a certain period of time. The computing system can use parcel data to identify parcels that are located at any one of the identified geographic locations. The computing system can further use the parcel data to identify a subset of the identified parcels that are listed for sale or rent. If the computing system identifies at least one parcel that the mobile device visited that is available on the market, this may indicate that the user who operates the mobile device may be shopping for a home. In response, the computing system can update information about the user to indicate that the user may be shopping for a home.

Abstract: This specification describes a simulation system that performs simulations of physical environments using a graph neural network. At each of one or more time steps in a sequence of time steps, the system can process a representation of a current state of the physical environment at the current time step using the graph neural network to generate a prediction of a next state of the physical environment at the next time step. Some implementations of the system are adapted for hardware GLOBAL acceleration. As well as performing simulations, the system can be used to predict physical quantities based on measured real-world data. Implementations of the system are differentiable and can also be used for design optimization, and for optimal control tasks.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating data specifying a three-dimensional mesh of an object using an auto-regressive neural network.

Abstract: A reinforcement learning system and method that selects actions to be performed by an agent interacting with an environment. The system uses a combination of reinforcement learning and a look ahead search: Reinforcement learning Q-values are used to guide the look ahead search and the search is used in turn to improve the Q-values. The system learns from a combination of real experience and simulated, model-based experience.

Abstract: A receptacle and a container formed therewith. The receptacle includes a bottom wall; a side wall and a rim extending laterally outwardly from the side wall. At least locking member is integrally formed with the rim and is selectively deflectable or oriented at an angle relative to an imaginary plane along which the rim extends. The locking members are a plurality of tabs that alternating with spaces defined the rim. In one embodiment, two identical receptacles form a container by inverting one receptacle relative to the other, deflecting the tabs through complementary spaces defined in the rim of the other receptacle, thereby interlocking the two receptacles together. In another embodiment, the receptacle includes a folding axis allowing first and second halves of the receptacle to be folded relative to each other. The tabs are deflected through complementary spaces defined in the opposing half, thereby interlocking the two halves together.

Abstract: A system implemented by one or more computers comprises a visual encoder component configured to receive as input data representing a sequence of image frames, in particular representing objects in a scene of the sequence, and to output a sequence of corresponding state codes, each state code comprising vectors, one for each of the objects. Each vector represents a respective position and velocity of its corresponding object. The system also comprises a dynamic predictor component configured to take as input a sequence of state codes, for example from the visual encoder, and predict a state code for a next unobserved frame. The system further comprises a state decoder component configured to convert the predicted state code, to a state, the state comprising a respective position and velocity vector for each object in the scene. This state may represent a predicted position and velocity vector for each of the objects.

Abstract: A neural network system is proposed. The neural network can be trained by model-based reinforcement learning to select actions to be performed by an agent interacting with an environment, to perform a task in an attempt to achieve a specified result. The system may comprise at least one imagination core which receives a current observation characterizing a current state of the environment, and optionally historical observations, and which includes a model of the environment. The imagination core may be configured to output trajectory data in response to the current observation, and/or historical observations. The trajectory data comprising a sequence of future features of the environment imagined by the imagination core. The system may also include a rollout encoder to encode the features, and an output stage to receive data derived from the rollout embedding and to output action policy data for identifying an action based on the current observation.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for predicting the actions of, or influences on, agents in environments with multiple agents, in particular for reinforcement learning. In one aspect, a relational forward model (RFM) system receives agent data representing agent actions for each of multiple agents and implements: an encoder graph neural network subsystem to process the agent data as graph data to provide encoded graph data, a recurrent graph neural network subsystem to process the encoded graph data to provide processed graph data, a decoder graph neural network subsystem to decode the processed graph data to provide decoded graph data and an output to provide representation data for node and/or edge attributes of the decoded graph data relating to a predicted action of one or more of the agents. A reinforcement learning system includes the RFM system.

Abstract: A system implemented by one or more computers comprises a visual encoder component configured to receive as input data representing a sequence of image frames, in particular representing objects in a scene of the sequence, and to output a sequence of corresponding state codes, each state code comprising vectors, one for each of the objects. Each vector represents a respective position and velocity of its corresponding object. The system also comprises a dynamic predictor component configured to take as input a sequence of state codes, for example from the visual encoder, and predict a state code for a next unobserved frame. The system further comprises a state decoder component configured to convert the predicted state code, to a state, the state comprising a respective position and velocity vector for each object in the scene. This state may represent a predicted position and velocity vector for each of the objects.

Abstract: A neural network system is proposed to select actions to be performed by an agent interacting with an environment to perform a task in an attempt to achieve a specified result. The system may include a controller to receive state data and context data, and to output action data. The system may also include an imagination module to receive the state and action data, and to output consequent state data. The system may also include a manager to receive the state data and the context data, and to output route data which defines whether the system is to execute an action or to imagine. The system may also include a memory to store the context data.

Abstract: A neural network system is proposed. The neural network can be trained by model-based reinforcement learning to select actions to be performed by an agent interacting with an environment, to perform a task in an attempt to achieve a specified result. The system may comprise at least one imagination core which receives a current observation characterizing a current state of the environment, and optionally historical observations, and which includes a model of the environment. The imagination core may be configured to output trajectory data in response to the current observation, and/or historical observations. The trajectory data comprising a sequence of future features of the environment imagined by the imagination core. The system may also include a rollout encoder to encode the features, and an output stage to receive data derived from the rollout embedding and to output action policy data for identifying an action based on the current observation.

Abstract: A graph neural network system implementing a learnable physics engine for understanding and controlling a physical system. The physical system is considered to be composed of bodies coupled by joints and is represented by static and dynamic graphs. A graph processing neural network processes an input graph e.g. the static and dynamic graphs, to provide an output graph, e.g. a predicted dynamic graph. The graph processing neural network is differentiable and may be used for control and/or reinforcement learning. The trained graph neural network system can be applied to physical systems with similar but new graph structures (zero-shot learning).

Abstract: A system implemented by one or more computers comprises a visual encoder component configured to receive as input data representing a sequence of image frames, in particular representing objects in a scene of the sequence, and to output a sequence of corresponding state codes, each state code comprising vectors, one for each of the objects. Each vector represents a respective position and velocity of its corresponding object. The system also comprises a dynamic predictor component configured to take as input a sequence of state codes, for example from the visual encoder, and predict a state code for a next unobserved frame. The system further comprises a state decoder component configured to convert the predicted state code, to a state, the state comprising a respective position and velocity vector for each object in the scene. This state may represent a predicted position and velocity vector for each of the objects.

Abstract: A neural network system is proposed to select actions to be performed by an agent interacting with an environment to perform a task in an attempt to achieve a specified result. The system may include a controller to receive state data and context data, and to output action data. The system may also include an imagination module to receive the state and action data, and to output consequent state data. The system may also include a manager to receive the state data and the context data, and to output route data which defines whether the system is to execute an action or to imagine. The system may also include a memory to store the context data.

Abstract: A neural network system is proposed. The neural network can be trained by model-based reinforcement learning to select actions to be performed by an agent interacting with an environment, to perform a task in an attempt to achieve a specified result. The system may comprise at least one imagination core which receives a current observation characterizing a current state of the environment, and optionally historical observations, and which includes a model of the environment. The imagination core may be configured to output trajectory data in response to the current observation, and/or historical observations. The trajectory data comprising a sequence of future features of the environment imagined by the imagination core. The system may also include a rollout encoder to encode the features, and an output stage to receive data derived from the rollout embedding and to output action policy data for identifying an action based on the current observation.

Abstract: A resealable cover for a beverage container that holds a carbonated beverage. The cover includes a base engageable with a rim of the container and a lid movable between an open and closed position relative to the base. When in the closed position, a passageway is defined between a portion of the lid and a portion of the base. The passageway permits fluid communication between air surrounding an exterior surface of the cover and a space defined between the lid and an opening in the container. Pressure that builds up under the lid from gas escaping from the liquid may be relieved by outgassing through the passageway. The cover includes latches on the base that engage in notches on the lid to help lock the lid in place. The lid is rotated relative to the base to release the latches and permit the lid to be opened.

Abstract: A brushing system with a rechargeable toothbrush and associated charging components include a charging cup, a charging base, and a charger. The charging cup is seated on the charging base which in turn is engaged with the charger. The charger includes a universal serial bus (USB) receptacle that is configured to receive a USB header extending outwardly from the charging base. The charging base includes a first charging coil in a protrusion that extends upwardly from the charging base and into a receptacle in the charging cup. The charging cup includes a projection that is received in a recess of the toothbrush. The toothbrush includes a second charging coil that is brought into the proximity of the first charging coil when the toothbrush is placed in the cup. The toothbrush is charged by induction when prongs on the charger are engaged in an electrical wall outlet.

Abstract: A neural network system is proposed. The neural network can be trained by model-based reinforcement learning to select actions to be performed by an agent interacting with an environment, to perform a task in an attempt to achieve a specified result. The system may comprise at least one imagination core which receives a current observation characterizing a current state of the environment, and optionally historical observations, and which includes a model of the environment. The imagination core may be configured to output trajectory data in response to the current observation, and/or historical observations. The trajectory data comprising a sequence of future features of the environment imagined by the imagination core. The system may also include a rollout encoder to encode the features, and an output stage to receive data derived from the rollout embedding and to output action policy data for identifying an action based on the current observation.

Abstract: A system implemented by one or more computers comprises a visual encoder component configured to receive as input data representing a sequence of image frames, in particular representing objects in a scene of the sequence, and to output a sequence of corresponding state codes, each state code comprising vectors, one for each of the objects. Each vector represents a respective position and velocity of its corresponding object. The system also comprises a dynamic predictor component configured to take as input a sequence of state codes, for example from the visual encoder, and predict a state code for a next unobserved frame. The system further comprises a state decoder component configured to convert the predicted state code, to a state, the state comprising a respective position and velocity vector for each object in the scene. This state may represent a predicted position and velocity vector for each of the objects.