Abstract: Embodiments of this application disclose a method and apparatus for training an image region segmentation model, and an image region segmentation method and apparatus. The method includes acquiring a sample image set, and each image of the sample image set having first annotation information; generating graph structure data corresponding to a sample image in the sample image set, the graph structure data comprising multiple nodes, and each node comprising at least one pixel in the sample image; determining second annotation information of each node according to the graph structure data and the first annotation information corresponding to the sample image by using a graph convolutional network model, a granularity of the second annotation information being smaller than a granularity of the first annotation information, the graph convolutional network model being a part of an image segmentation model; and training the image segmentation model according to the second annotation information.

Abstract: The present invention provides an electric power material distribution and allocation method based on an optimized branch and bound method, including: sorting the distribution of each batch of electric power materials to obtain a material set R; screening available vehicle data to obtain a vehicle deadweight set W and a vehicle unique identification set C; determining whether available vehicles meet requirements; and using an optimized queued branch and bound method to obtain an optimal distribution and allocation scheme. According to the present invention, scientific scheme support is provided for the distribution and allocation of the electric power materials, the scientificity of the distribution of power grid materials is improved, the influence of human factors is reduced, the utilization rate of distribution vehicles is increased, the distribution cost of the materials is reduced, and the economic benefit is improved.

Abstract: For a set in sets of candidate factors, a classification model is trained to predict a computer problem possibility, an accuracy score of the classification model is determined based on model validation, and factor weights of the candidate factors in the set are adjusted based on the accuracy score. This processing is done with respect to all sets of candidate factors. A low accuracy classification model having an accuracy score lower than a threshold criterion is selected. A higher accuracy classification model having an accuracy score that is higher than the accuracy score of the low accuracy classification model is selected. The set of candidate factors used to train the low accuracy classification model is updated using one or more of the candidate factors used to train the higher accuracy classification model. The low accuracy classification model is updated based on the updated set of candidate factors.

Abstract: Embodiments of this application relate to control parameter optimization technologies. Various embodiments include methods and apparatus for optimizing control parameters. At least one group of initial control parameters within a preset range is randomly generated, to obtain a group of current control parameters; it is determined whether a number of updates has exceeded a preset threshold, referring to a total number of times that the current control parameters have been changed; the group of current control parameters is changed when the number of updates does not reach the preset threshold; and air conditioning control parameters are optimized according to control parameters corresponding to a comprehensive score of a combination when the number of updates reaches the preset threshold.

Abstract: Computer implemented methods, systems, and computer program products include program code executing on a processor(s) obtains a given query referencing a LOB. The program code obtains statistics and activities performed in the database system, establishes performance benchmarks, and analyzes the given query to determine how the LOB is utilized. The program code determines that meeting a benchmark requires replacing a reference to the LOB with a reference to implicit column(s). The program code executes the given query by referencing the implicit column(s) to obtain query results. The program code returns the query results.

Abstract: A light-emitting diode substrate, a manufacturing method thereof, and a display device are disclosed. The manufacturing method of the light-emitting diode substrate includes: forming an epitaxial layer group of M light-emitting diode chips on a substrate; transferring N epitaxial layer groups on N substrates onto a transition carrier substrate, the N epitaxial layer groups on the N substrates being densely arranged on the transition carrier substrate; and transferring at least part of N*M light-emitting diode chips corresponding to the N epitaxial layer groups on the transition carrier substrate onto a driving substrate, an area of the transition carrier substrate is greater than or equal to a sum of areas of the N substrates, M is a positive integer greater than or equal to 2, and N is a positive integer greater than or equal to 2.

Abstract: A computer-implemented method, system and computer program product for improving accuracy and efficiency of auditing databases. A table, list or index of a database is analyzed to identify metadata, which includes time series data, user data, an Internet Protocol address and operation data. The identified metadata is associated with the corresponding record or row of the table, list or index from which the metadata was extracted. A determination is then made as to whether to record a raw data image associated with the record or row of the analyzed table, list or index based on the corresponding data operation. The identified metadata as well as the recorded data images, if any, are stored in a structured audit log. Auditing information is then obtained from a structured audit log based on matching the record or row identifier (RID) associated with the query with the RID associated with the structured audit log.

Abstract: Embodiments analyze a query pattern of an incoming query on a database, perform a semantic analysis of the query pattern of the incoming query, generate a re-write query that has an improved query performance in comparison to a query performance of the incoming query based on the analyzed query pattern and the semantic analysis; build a query model using machine learning based on at least one of the query pattern and the semantic analysis; and apply the re-write query by performing the re-write query on the database to provide the improved query performance.

Abstract: The invention provides methods, devices and non-transitory tangible computer-readable storage media for estimating a state of charge of a battery and extracting a charging curve of the battery. The method for estimating the state of charge of the battery includes estimating a first state of charge of the battery; determining a charging curve that conforms to the present charging scenario, wherein the present charging scenario is characterized by at least one of parameters including an ambient temperature, a charging rate, an internal resistance, a state of health, a lifespan, and an open circuit voltage of the battery; and correcting the first state of charge according to the charging curve to obtain a final state of charge of the battery.

Abstract: An array substrate has a display area and a bonding region. The display area includes a distal region, a proximal region, and a middle region therebetween. The array substrate includes a base, a common electrode located in the display area, a connecting lead disposed outside the distal region, a conductive frame at least partially surrounding the display area, and at least one first common signal line, at least one second common signal line and at least one third common signal line. The first common signal line, the second common signal line and the third common signal line are respectively coupled to portions of the common electrode located in the distal region, the proximal region and the middle region. The first common signal line is coupled to the connecting lead. The connecting lead and the portion of the common electrode located in the distal region are coupled to the conductive frame.

Abstract: A recombinant hybrid protein, a nucleic acid molecule encoding the recombinant hybrid protein, a transgenic expression cassette containing the nucleic acid molecule, and a gene delivery system containing the transgenic expression cassettes that can be used for treating muscular dystrophy. The nucleic acid molecule, the transgenic expression cassette and the gene delivery system can stably express a therapeutic protein in muscle tissues, so that a good therapeutic effect on muscular dystrophy is realized.

Abstract: In several aspects, a computing device analyzes data to determine its characteristics. The computing device selects at least one compression process based on the characteristics. A compression switch and a compression level are dynamically adjusted based on multiple factors including available system resources, desired storage savings and performance requirement. The compression level dynamically varies depending on the data being processed and a workload on a system. Performance of compression and decompression operations are continuously monitored for dynamically adjusting compression parameters to optimize performance.

Abstract: A computer-implemented method, including receiving, by a processor set, a query including a query string for a system catalog; identifying, by the processor set, a default index structure of the system catalog; executing, by the processor set, the query based on the default index structure of each index in the system catalog; ranking, by the processor set, a performance of each execution of the query by each index; mapping, by the processor set, a query pattern to a corresponding index of the system catalog; selecting, by the processor set, the index to perform the query using a machine learning (ML) model trained with a knowledge base that includes the ranking and the mapping; executing, by the processor set, the query on the selected index; and in response to executing the query on the selected index, returning, by the processor set, a result of the query.

Abstract: A bone tissue regeneration product, a method for preparing the bone tissue regeneration product and, and a method for repairing a bone defect. The new use of IGSF10 of the present disclosure can reduce the effective concentration of BMP without obvious adverse reactions, thereby exploring a method to promote bone tissue regeneration and providing new ideas for the treatment of bone defects. The present disclosure is capable of promoting non-bone related cells to produce bones; the method of the present disclosure is also capable of inhibiting the bone loss in inflammatory environments.

Abstract: The present disclosure provides an array substrate and a display device. The array substrate includes: a plurality of gate lines extending in a first direction, and a plurality of data lines extending in a second direction and crossing the gate lines to define a plurality of sub-pixels; a plurality of touch signal lines extending in the second direction and arranged in light shielding regions of the sub-pixels; a plurality of touch electrodes insulated from each other; and a plurality of metal pattern units corresponding to the sub-pixels respectively and arranged in the light shielding region of each sub-pixel. The metal pattern unit includes a first metal strip arranged on at least one side of the data line and extending in the second direction.

Abstract: A method, an apparatus and a device for identifying battery parameters, and a storage medium are provided.

Abstract: Provided is a fully decoupled MEMS gyroscope, including an anchor point unit, a sensing unit elastically connected to the anchor point unit, and a driving unit configured to drive the sensing unit to move. The anchor point unit includes a center anchor point subunit located at a center of a rectangle and four side anchor points. The driving unit includes four driving members located on four sides of the rectangle. The sensing unit includes two X mass blocks symmetrically arranged in two avoiding intervals, two Y mass blocks symmetrically arranged in the other two avoiding intervals, four Z mass blocks respectively located at an outer side of each driving member, and four Z detection decoupling members respectively located at an outer side of each Z mass block. The X mass blocks and the Y mass blocks are respectively connected to each side anchor point.

Abstract: A display panel is provided, which includes a first substrate and a second substrate arranged opposite to each other. The first substrate includes a first electrode and a second electrode. The second substrate includes a light shielding layer. The light shielding layer includes a light transmitting region and a light shielding region. The first electrode includes slits extending in a first direction, and an orthographic projection of two ends of at least one of the slits onto the first substrate is within an orthographic projection of the light shielding region onto the first substrate. A display panel and a display device are also provided.

Abstract: The illustrative embodiments provide for dynamic tuning of pre-initialization environment provisioning and management. An embodiment includes accepting a request from a group of applications to generate a performance-based index table for a workload based on a feature of the applications and generating the performance-based index table. The embodiment includes building a label feature by analyzing a static program feature of the applications and the performance-based index table. The embodiment includes constructing, using clustering algorithms, a model for provisioning a pre-initialization environment using the label features. The embodiment includes loading the applications into a pre-initialization environment. The embodiment includes introducing a selection policy for a switch in the pre-initialization environment in multiple applications to balance usage of a resource.

Abstract: A system and method for improving the performance and reducing costs of a program by automatically provisioning and managing proper memory pool cell size adaptive to each executing application. By collecting time series of historical data on the memory pool usage of applications over a period of time, respective time-series prediction models are used to process the data to predict the allocation size for applications and in particular, a predicted number of allocations and a respective predicted allocation cell size. A clustering-based method is further applied to predict the allocation size for applications, using real time execution to do scaling, complement and interpolation. A method runs a further time-series prediction model trained to predict, based on the predicted memory cell size and one or more application profile features associated with the requesting application, a tuning parameter to refine the memory pool storage area size used for handling memory allocation requests.