Abstract: The disclosure provides a monitoring system for external environment of an oil well platform, and relates to the technical field of real-time monitoring of external environment of the oil well platform, and includes a sensor unit, a satellite transmission unit, a data collector, a central server and a control unit, where sensors are used for collecting detection data and sending the detection data to the satellite transmission unit; the satellite transmission unit stores the detection data to a local storage device in real time through edge calculation and transmits the detection data to a base; the base receives the detection data and then sends the detection data to the data collector; and the data collector receives the detection data and transmits the detection data to the central server.

Abstract: The present invention provides a model training system, a model training method, a training device, and a training node. The system includes: n training devices, in which a total of M training nodes are deployed. Each training device stores a global index list and a dataset corresponding to a training node deployed in the training device, and the global index list indicates indexes of data in datasets respectively corresponding to the M training nodes. Each training device is configured to: shuffle the global index list by using a random number seed, and query, based on a shuffled list, the dataset stored in the training device, to obtain exchange data of the training node deployed in the training device, where the exchange data is hit data obtained through data query.

Abstract: A data processing method is provided, applied to the field of artificial intelligence. The method includes: determining, based on a target mapping relationship, one or more target network units corresponding to a target word vector and a storage address, where storage space corresponding to the storage address is located in storage space outside a computing unit; obtaining the one or more target network units from the storage space; and performing, through the computing unit based on the target word vector, a training process corresponding to a neural network constructed based on the one or more target network units. Because storage space of the storage location outside the computing unit may be set to be relatively large, through separation of storage and compute, a size of the large-scale model during training can be increased and scalability and flexibility of the large-scale model can be improved.

Abstract: The disclosure provides a monitoring system for external environment of an oil well platform, and relates to the technical field of real-time monitoring of external environment of the oil well platform, and includes a sensor unit, a satellite transmission unit, a data collector, a central server and a control unit, where sensors are used for collecting detection data and sending the detection data to the satellite transmission unit; the satellite transmission unit stores the detection data to a local storage device in real time through edge calculation and transmits the detection data to a base; the base receives the detection data and then sends the detection data to the data collector; and the data collector receives the detection data and transmits the detection data to the central server.

Abstract: The present disclosure describes, in part, an enzyme-mediated radical initiating system and methods of using the system to produce polymers, including polymeric hydrogels, at ambient conditions.

Abstract: Provided is an intelligent monitoring system for an oil development platform. The system includes external environment monitoring of a well platform, internal environment monitoring of the well platform and platform construction, where the external environment monitoring of the well platform includes real-time monitoring of natural environment conditions around an offshore oil development platform, the internal environment monitoring of the well platform includes collection and analysis of various indexes inside the oil well platform, and the platform construction is based on a three-layer structural design including a sensing layer, a transmission layer and an application layer. The sensing layer collects environmental data of ocean and the well platform through an Internet of Things technology, the transmission layer rapidly transmits data to ensure real-time updating of information, and the application layer processes received data to realize fusion and visualization of the data.

Abstract: The disclosure relates to a technical field of oil monitoring, and in particular to real-time monitoring system for an internal environment of an oil well platform. The system includes: a group of Internet of Things sensors used for collecting data of the internal environment of the oil well platform in real time; a satellite transmission unit communicated with a plurality of Internet of Things sensors and used for receiving data collected by the Internet of Things sensors; the satellite receiving terminal used for receiving data from the satellite transmission unit and transmitting data to a central server; the central server with a data processing function used for receiving data from the satellite receiving terminal and carrying out data integration and analysis; and a control center used for displaying monitoring results and data trends from the central server, and issuing early warning or alarm based on a preset environmental parameter standard.

Abstract: The present invention provides a model training system, a model training method, a training device, and a training node. The system includes: n training devices, in which a total of M training nodes are deployed. Each training device stores a global index list and a dataset corresponding to a training node deployed in the training device, and the global index list indicates indexes of data in datasets respectively corresponding to the M training nodes. Each training device is configured to: shuffle the global index list by using a random number seed, and query, based on a shuffled list, the dataset stored in the training device, to obtain exchange data of the training node deployed in the training device, where the exchange data is hit data obtained through data query.

Abstract: The present disclosure describes, in part, an enzyme-mediated radical initiating system and methods of using the system to produce polymers, including polymeric hydrogels, at ambient conditions.

Abstract: Embodiments of this application disclose a data processing method, and relate to the field of artificial intelligence. The method is applied to distributed parallel model training, for example, distributed training of a text translation model, a speech recognition model, a facial recognition model, a three-dimensional reconstruction model, and a virtual reality model. The method can implement hybrid parallelism in a distributed cluster. The method includes: inserting, based on tensor layouts of tensors of at least one operator in a deep neural network model, a redistribution operator between operators that have an input-output dependency relationship, to implement conversion between different tensor layouts; inserting the redistribution operator into a sliced computational graph; and determining an updated sliced computational graph to implement parallel model training of the deep neural network.

Abstract: The present disclosure describes, in part, an enzyme-mediated radical initiating system and methods of using the system to produce polymers, including polymeric hydrogels, at ambient conditions.

Abstract: A memory device and a data reading method are provided. A dummy circuit performs a read operation in synchronism with a data access circuit according to an address signal, so as to estimate time points at which the data access circuit completes each of operating procedures, and enable the data access circuit to execute a next operating procedure when completing an operating procedure.

Abstract: Any number of Serial Peripheral Interface (“SPI”) flash memory die may be stacked and packaged using any desired multi-chip packaging technique to realize any one or combination of various capabilities such as low per-bit cost, high density storage, code shadowing to RAM, and fast random access for “execute in place” applications, while preserving the advantages of the SPI interface. During device manufacture, each of the stacked die is assigned a unique identifier or “Die ID” relative to the other stacked die in the package. During normal operations, the unique Die IDs are used by a Die Select instruction to enable one of the stacked die to respond to subsequent instructions on the SPI interface, while disabling the other stacked die in the package from responding to subsequent instructions but for certain “Universal” instructions which include the Die Select instruction. Concurrent operations by the stacked die are supported.

Abstract: Any number of Serial Peripheral Interface (“SPI”) flash memory die may be stacked and packaged using any desired multi-chip packaging technique to realize any one or combination of various capabilities such as low per-bit cost, high density storage, code shadowing to RAM, and fast random access for “execute in place” applications, while preserving the advantages of the SPI interface. During device manufacture, each of the stacked die is assigned a unique identifier or “Die ID” relative to the other stacked die in the package. During normal operations, the unique Die IDs are used by a Die Select instruction to enable one of the stacked die to respond to subsequent instructions on the SPI interface, while disabling the other stacked die in the package from responding to subsequent instructions but for certain “Universal” instructions which include the Die Select instruction. Concurrent operations by the stacked die are supported.

Abstract: A method includes steps of: providing a first memory cell array including a plurality of first word lines, wherein a plurality of first data are stored in the first memory cell array; providing a second memory cell array including a plurality of second word lines, wherein the second memory cell array is separated from the first memory cell array, and a plurality of second data are stored in the second memory cell array; selecting one of the first word lines and one of the second word lines at a same time or an overlapping time; alternately selecting a first address of the first memory cell array and a second address of the second memory cell array to alternately read a first corresponding portion of the first data and a second corresponding portion of the second data from the first memory cell array and the second memory cell array.

Abstract: A method of erasing a target erase area of a non-volatile memory is provided, wherein the non-volatile memory is divided into an target erase area and an unselected area, and the method includes the steps in an erase cycle of: conditioning the target erase area of the non-volatile memory, wherein the unselected area is an area, excluding the target erase area, in the non-volatile memory; erasing target cells of the target erase area, wherein the threshold of the target cells is not greater than an erase verify voltage; soft-programming the target cells, wherein the threshold of the target cells is not less than a soft program verify voltage, wherein the soft program verify voltage is less than the erase verify voltage; and refreshing a predefined portion of the unselected area, wherein the predefined portion in the erase cycle is less than the unselected area.

Abstract: A method includes steps of: providing a first memory cell array including a plurality of first word lines, wherein a plurality of first data are stored in the first memory cell array; providing a second memory cell array including a plurality of second word lines, wherein the second memory cell array is separated from the first memory cell array, and a plurality of second data are stored in the second memory cell array; selecting one of the first word lines and one of the second word lines at a same time or an overlapping time; alternately selecting a first address of the first memory cell array and a second address of the second memory cell array to alternately read a first corresponding portion of the first data and a second corresponding portion of the second data from the first memory cell array and the second memory cell array.

Abstract: A data storage device having a non-volatile memory and a controller and a control method for the non-volatile memory are disclosed. The non-volatile memory has a plurality of blocks for data storage and each block provides a plurality of sectors. The controller allocates erase marker bits in each of the sectors to record the progress of an erase operation performed on the non-volatile memory for resumption of the erase operation when required.

Abstract: A data storage device having a non-volatile memory and a controller and a control method for the non-volatile memory are disclosed. The non-volatile memory has a plurality of blocks for data storage and each block provides a plurality of sectors. The controller allocates erase marker bits in each of the sectors to record the progress of an erase operation performed on the non-volatile memory for resumption of the erase operation when required.

Abstract: The timing of logic read operations in a Flash memory device may be improved by a pad serial output circuit which receives a pre-decoded instruction signal and pre-fetched logic data prior to the last command clock, and which performs a fast resolution of the command in the pad serial output circuit on the last clock of the command input sequence. In one illustratively implementation, instruction pre-decode and data pre-fetch may be done on the seventh clock during command input. In another illustrative implementation, a first instruction pre-decode and data pre-fetch may be done on the fourth clock during command input, and a second instruction pre-decode may be done on the seventh clock during command input. Both serial protocol interface, including dual and quad I/O SPI, and quad peripheral interface are supported.