Abstract: Systems and methods are provided for generating a statistical parameter representing a state of a surgical procedure from sensor data. Sensor data representing a time period. is received from a sensor. Numerical features representing the time period are generated from the sensor data. Each of a plurality of long short term memory units are updated according to the plurality of numerical features via a message passing process. The long short term memory units are connected to form a graph, with a first set of the long short term memory units representing a plurality of nodes of the graph and a second set of the long short term memory units representing a plurality of hyperedges of the graph. A statistical parameter representing a state of the surgical procedure for the time period is derived from an output of one of the long short term memory units and provided to a user.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: This disclosure describes systems, methods, and apparatus for a locking folding knife having a handle lock and a safety switch arranged around and rotatable around a pivot axis of the knife. In an engaged position the switch overlaps at least a portion of the handle lock and prevents it from moving into an unlocked position. A spring, such as a wire form spring, can be arranged within a pocket of the switch, and when the switch is moved between engaged and disengaged positions the spring can be compressed generating a rotational bias on the switch urging the switch toward either the engaged or disengaged position.

Abstract: A castor might comprise a castor body for coupling the castor to an apparatus, at least one wheel coupled to the castor body, wherein said wheel is operable to rotate about a wheel axis. The castor body might be configured to swivel about a swivel axis which is substantially perpendicular to the wheel axis such as to rotate the orientation of the wheel with respect to the swivel axis. A locking means might be configured to adopt at least first state and a second state. In the first state, the locking means might be configured to allow swivelling motion of the castor body about the swivel axis and in the second state the locking means might be configured to inhibit the swivelling motion of the castor body about the swivel axis such that the orientation of the castor wheel is locked relative to the swivel axis.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Systems and methods are provided for generating a statistical parameter representing a state of a surgical procedure from sensor data. Sensor data representing a time period. is received from a sensor. Numerical features representing the time period are generated from the sensor data. Each of a plurality of long short term memory units are updated according to the plurality of numerical features via a message passing process. The long short term memory units are connected to form a graph, with a first set of the long short term memory units representing a plurality of nodes of the graph and a second set of the long short term memory units representing a plurality of hyperedges of the graph. A statistical parameter representing a state of the surgical procedure for the time period is derived from an output of one of the long short term memory units and provided to a user.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Embodiments of a fluid warming container and a fluid warming device are described. The fluid warming container defines an interior volume arranged to receive a liquid. The fluid warming container also includes a heating element, a power source, and a switch configured to selectively control a flow of electrical current from the power source to the heating element to warm the liquid. The fluid warming device includes an elongate housing, a heating element, a power source, and a switch configured to selectively control a flow of electrical current from the power source to the heating element to warm a liquid in contact with the fluid warming device.

Abstract: A combination radar arch and bimini top includes a radar arch securable on a boat deck over a boat cabin and a bimini top assembly secured to the radar arch. The bimini top assembly includes at least one bimini top base secured on the radar arch, a bimini arm pivotally connected to the bimini top base, and a spring-loaded roller assembly secured to the radar arch and including a canopy connected between the radar arch and the bimini arm. The bimini top base has a drive mechanism coupled with the bimini arm that is configured to displace the bimini arm between a retracted position and an extended position.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Visualizing event impact risk over time. At least one location of interest is identified on a geospatial map, and an event path is plotted on the geospatial map for an event over time. The event path includes actual and/or forecasted paths of the event. Along the event path, connected geospatial polygons are generated. The geospatial polygons represent actual and/or forecasted intensity of the event over time. A determination is made as to whether the connected geospatial polygons intersect one or more of the locations of interest indicating an impact risk for the event. Action are taken to reduce the impact risk for the location of interest.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Systems, methods, and apparatus are described for a locking folding knife having a handle lock and a safety switch arranged around and rotatable around a pivot axis of the knife. In an engaged position the switch overlaps at least a portion of the handle lock and prevents it from moving into an unlocked position. A spring, such as a wire form spring, can be arranged within a pocket of the switch, and when the switch is moved between engaged and disengaged positions the spring can be compressed generating a rotational bias on the switch urging the switch toward either the engaged or disengaged position.

Abstract: A shade system includes a first linear actuator attachable to a structure, and a second linear actuator spaced from the first linear actuator and attachable to the structure. The first and second linear actuators are each displaceable between a retracted position and an extended position. A canvas connected between the first and second linear actuators is fixed to distal ends and to proximal ends of the first and second linear actuators and is free-floating between the distal and proximal ends. The canvas is configured to retract into an accordion configuration when the first and second linear actuators are displaced from the extended position to the retracted position.

Abstract: Methods, systems, and computer program products for identifying and evaluating risk alerts are provided. Aspects include receiving risk alert data comprising risk alerts from data sources, obtaining, by the processor, client geographical data comprising one or more locations of interest, wherein the one or more locations of interest comprise infrastructure resource for the client, determining one or more trusted data sources that are most relevant to define an impact of one or more risk events from the risk alert data affecting one or more locations of interest for the client based on the risk alert data and the client geographical data, determining impactful risk events that affect locations of interest for the client based on trusted risk alert data from the trusted data sources and the client geographical data, and displaying the trusted data sources, the one or more impactful risk events, and the one or more locations.

Abstract: This disclosure describes systems, methods, and apparatus for a locking folding knife having a handle lock and a safety switch arranged around and rotatable around a pivot axis of the knife. In an engaged position the switch overlaps at least a portion of the handle lock and prevents it from moving into an unlocked position. A spring, such as a wire form spring, can be arranged within a pocket of the switch, and when the switch is moved between engaged and disengaged positions the spring can be compressed generating a rotational bias on the switch urging the switch toward either the engaged or disengaged position.

Abstract: A combination radar arch and bimini top includes a radar arch securable on a boat deck over a boat cabin and a bimini top assembly secured to the radar arch. The bimini top assembly includes at least one bimini top base secured on the radar arch, a bimini arm pivotally connected to the bimini top base, and a spring-loaded roller assembly secured to the radar arch and including a canopy connected between the radar arch and the bimini arm. The bimini top base has a drive mechanism coupled with the bimini arm that is configured to displace the bimini arm between a retracted position and an extended position.

Abstract: This disclosure describes systems, methods, and apparatus for a locking folding knife having a handle lock and a safety switch arranged around and rotatable around a pivot axis of the knife. In an engaged position the switch overlaps at least a portion of the handle lock and prevents it from moving into an unlocked position. A spring, such as a wire form spring, can be arranged within a pocket of the switch, and when the switch is moved between engaged and disengaged positions the spring can be compressed generating a rotational bias on the switch urging the switch toward either the engaged or disengaged position.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.