Abstract: A water heater includes power distribution circuitry receiving primary power and secondary power. The water heater includes a heat exchange system including at least one component powered via the primary power and configured to transfer heat between a refrigerant and water. The water heater includes at least one heater operably coupled with and configured to heat at least one portion of the heat exchange system. The water heater includes switching circuitry electrically interposing the power distribution circuitry and the at least one heater to selectively power the at least one heater via the primary power in a first condition of the switching circuitry and via the secondary power when the primary power is unavailable in a second condition of the switching circuitry.

Abstract: Systems and methods for language-agnostic unit testing are generally described. In various examples, first data including first source code written in a first programming language may be received. Configuration data configured to execute language-agnostic unit testing may be loaded. Input data may be determined from the first data based on a first input defined by the configuration data for a first unit test. A first driver written in the first programming language may be determined from the configuration data. The first driver may include computer-executable instructions configured to execute the first unit test. First output data may be generated by executing the first computer-executable instructions to perform the first unit test on the first input data.

Abstract: Systems and methods for generating asynchronous application programming interface (API) documents are generally described. In various examples, a topic discovery request may be sent to a message broker. A list of topics may be received from the message broker. First identifier data identifying a first topic of the list of topics may be received from the message broker. A first message may be received from the message broker, where the first message pertains to the first topic. First schema may be determined using the first message. The first schema data may describe content of the first message and organization of the content of the first message. The first identifier data may be stored in a first data structure in association with the first schema data.

Abstract: A method of training a reinforcement machine learning computer system. The method comprises providing a machine-learning computer programming language including a pre-defined plurality of reinforcement machine learning criterion statements, and receiving a training specification authored in the machine-learning computer programming language. The training specification defines a plurality of training sub-goals with a corresponding plurality of the reinforcement machine learning criterion statements supported by the machine-learning computer programming language. The method further comprises computer translating the plurality of training sub-goals from the training specification into a shaped reward function configured to score a reinforcement machine learning model configuration with regard to the plurality of training sub-goals.

Abstract: A method of training a machine learning system. The method comprises collecting a first simulation dataset derived from a computer simulating a hypothetical scenario with a first simulation configuration having a first degree of fidelity. The method further comprises collecting a second simulation dataset derived from a computer simulating the hypothetical scenario with a second simulation configuration having a second degree of fidelity different than the first degree of fidelity. The method further comprises building a multi-fidelity training dataset including training data from both the first simulation dataset and the second simulation dataset according to an interleaving protocol.

Abstract: Systems and methods for generating asynchronous application programming interface (API) documents are generally described. In various examples, a topic discovery request may be sent to a message broker. A list of topics may be received from the message broker. First identifier data identifying a first topic of the list of topics may be received from the message broker. A first message may be received from the message broker, where the first message pertains to the first topic. First schema may be determined using the first message. The first schema data may describe content of the first message and organization of the content of the first message. The first identifier data may be stored in a first data structure in association with the first schema data.

Abstract: Systems and methods for scalable collaborative editing are generally described. In various examples, a first collaborative editing session may be executed at a first node of a first cluster of compute nodes. A load balancer executing on the first cluster of compute nodes may receive a first request from a first computing device. In various examples, the first request may include first identifier data identifying the first collaborative editing session. In some examples, the load balancer may determine that the first identifier data is associated with the first collaborative editing session. Further, the load balancer may determine that the first collaborative editing session is assigned to the first node. In some examples, the first request may be sent to the first node.

Abstract: Systems and methods for scalable collaborative editing are generally described. In various examples, a first collaborative editing session may be executed at a first node of a first cluster of compute nodes. A load balancer executing on the first cluster of compute nodes may receive a first request from a first computing device. In various examples, the first request may include first identifier data identifying the first collaborative editing session. In some examples, the load balancer may determine that the first identifier data is associated with the first collaborative editing session. Further, the load balancer may determine that the first collaborative editing session is assigned to the first node. In some examples, the first request may be sent to the first node.

Abstract: Systems and methods for generating asynchronous application programming interface (API) documents are generally described. In various examples, a metadata discovery request may be sent to a message broker. A list of topics may be received from the message broker. First metadata related to a first topic of the list of topics may be received from the message broker. A first message may be received from the message broker, where the first message pertains to the first topic. A first schema may be determined using the first message pertaining to the first topic. The first schema may include data describing content of the first message and organization of the content of the first message. An asynchronous API document may be generated based at least in part on the first schema and the first metadata.

Abstract: Systems and methods for generating asynchronous application programming interface (API) documents are generally described. In various examples, a metadata discovery request may be sent to a message broker. A list of topics may be received from the message broker. First metadata related to a first topic of the list of topics may be received from the message broker. A first message may be received from the message broker, where the first message pertains to the first topic. A first schema may be determined using the first message pertaining to the first topic. The first schema may include data describing content of the first message and organization of the content of the first message. An asynchronous API document may be generated based at least in part on the first schema and the first metadata.

Abstract: A method of training a machine learning system. The method comprises collecting a first simulation dataset derived from a computer simulating a hypothetical scenario with a first simulation configuration having a first degree of fidelity. The method further comprises collecting a second simulation dataset derived from a computer simulating the hypothetical scenario with a second simulation configuration having a second degree of fidelity different than the first degree of fidelity. The method further comprises building a multi-fidelity training dataset including training data from both the first simulation dataset and the second simulation dataset according to an interleaving protocol.

Abstract: A method of training a reinforcement machine learning computer system. The method comprises providing a machine-learning computer programming language including a pre-defined plurality of reinforcement machine learning criterion statements, and receiving a training specification authored in the machine-learning computer programming language. The training specification defines a plurality of training sub-goals with a corresponding plurality of the reinforcement machine learning criterion statements supported by the machine-learning computer programming language. The method further comprises computer translating the plurality of training sub-goals from the training specification into a shaped reward function configured to score a reinforcement machine learning model configuration with regard to the plurality of training sub-goals.

Abstract: A hydrostatic transmission for a working machine having a frame, an engine on the frame, and at least one wheel driven by the hydrostatic transmission is disclosed. The hydrostatic transmission includes a housing, a hydraulic pump, the hydraulic pump driven by an output shaft from the engine, a hydraulic motor for driving an output shaft connected to the at least one wheel, a first fluid line fillable with hydraulic fluid to be maintained under a first pressure, a second fluid line fillable with hydraulic fluid to be maintained under a second pressure, a first check valve for allowing hydraulic fluid to enter the first fluid line, a second check valve for allowing hydraulic fluid to enter the second fluid line, and an entrained air deflector for preventing hydraulic fluid entrained with air from entering the first check valve and the second check valve.

Abstract: A method for irradiating skin and deeper tissue using unpackaged LED lights to improve skin and tissue absorption of low level light energy may be provided. The apparatus embodied by the method may be formed by selectively mounting chip-on-board LEDs in various array so that the skin gets irradiated from all angles. The LEDs may be associated with a heat sink provided by the substrate of PCB to more effectively dissipate its heat, allowing the apparatus to run at higher outputs.

Abstract: A hydrostatic transmission for a working machine having a frame, an engine on the frame, and at least one wheel driven by the hydrostatic transmission is disclosed. The hydrostatic transmission includes a housing, a hydraulic pump, the hydraulic pump driven by an output shaft from the engine, a hydraulic motor for driving an output shaft connected to the at least one wheel, a first fluid line fillable with hydraulic fluid to be maintained under a first pressure, a second fluid line fillable with hydraulic fluid to be maintained under a second pressure, a first check valve for allowing hydraulic fluid to enter the first fluid line, a second check valve for allowing hydraulic fluid to enter the second fluid line, and an entrained air deflector for preventing hydraulic fluid entrained with air from entering the first check valve and the second check valve.

Abstract: The present invention provides a computer implemented method for communicating a signature in a received email. Initially, a server receives a source email from a source client, the source email comprising a signature index entry, and a body. The server determines for a first recipient that a first associated destination client is capable of rendering a signature of the signature index entry in response to a selection of an indexed signature. The server, responsive to a determination that the destination client is capable of rendering the signature index entry, forwards the source email to the first recipient as a delivered email using the body of the source email such that an indexed signature is present in the body of the delivered email. In addition, a method for generating a signature header is shown that provides a many to one correspondence between indexed signatures and a signature index entry.

Abstract: An improved door assembly includes a door frame having a door jamb and a door having a fixed portion including a hinged side and a locking side where the hinged side of the door is hinged to the door frame. At least one of the hinged side and the locking side of the door has at least one pivotable door portion configured to pivotably move between a retracted position inside the door and an extended position that extends inside at least one complementary shaped channel within the door jamb of the door frame to interlock the door and the door frame.

Abstract: A computer implemented method, apparatus, and computer usable program code for managing a software partition in logical partitioned data processing system. A virtual root process is created for the software partition, wherein the virtual root process functions as a root process for the software partition. The set of child processes is re-parented to the virtual root process in response to termination of a process having a set of child processes.

Abstract: A computer-implemented method for managing messages. Responsive to receiving a message at a data processing system, wherein the message comprises only a number of message identifiers and metadata identifying the message as pointing to a number of messages in a message history stored in the data processing system, the number of messages in the message history on the data processing system is identified to form an identified number of messages. A presence of the identified number of messages in the message history is displayed on a display in the data processing system.