Abstract: A capillary array includes a capillary region, including capillaries of a first glass, which are disposed in an axis-parallel manner. A low refractive index layer is disposed on an inner wall of each of the capillaries, the refractive index of each low refractive index layer being less than a refractive index of a liquid scintillator. A second glass material is disposed between adjacent capillaries. A softening point of the first glass is T1, a softening point of second glass is T2, and a value of T1 minus T2 is in a range from 30° C. to 50° C. A thermal expansion coefficient of the first glass is ?1. An edge covering region is disposed on an outer side of the capillary region and makes contact with an outer side face of the capillary region, wherein a material of the edge covering region is a third glass.

Abstract: This application discloses a communication method and apparatus. The communication method and apparatus are applicable to an Internet of Vehicles, or are applicable to fields such as D2D, intelligent driving, and intelligent network connected vehicles.

Abstract: The present disclosure relates to a method for screening an acetylcholinesterase inhibitor, and further relates to a kit and system for screening an acetylcholinesterase inhibitor.

Abstract: A method of configuring a mark having a trench to be etched into a substrate, the method including: obtaining a relation between an extent of height variation across a surface of a probationary layer deposited on a probationary trench of a probationary depth and a thickness of the probationary layer; determining an extent of height variation across the surface of a layer deposited on the mark allowing a metrology system to determine a position of the mark; and configuring the mark by determining a depth of the trench based on the relation, the extent of height variation and the thickness of a process layer to be deposited on the mark.

Abstract: A hybridoma cell line of secreting cyproheptadine monoclonal antibodies with a preservation number of CGMCC No. 14699 belongs to the field of food safety immunological detection. BALB/c mice are immunized through one time immunization with complete freund's adjuvant, three times of booster immunization with incomplete freund's adjuvant and one time of rush immunization with cyproheptadine complete antigen without adjuvant; the spleen cells from BALB/C mice immunized with high potency and low value of IC50 are fused with murine myeloma cells; and then the hybridoma cell line is obtained through indirect competitive ELISA screening and three sub-clones. The monoclonal antibody secreted by this cell line has good specificity and detection sensitivity to cyproheptadine (value of IC50 is 0.37 ng/ml), being suitable for detection of cyproheptadine in food.

Abstract: Embodiments of the present disclosure provide a method and apparatus for executing program service logic, a device and a medium. The method includes: dividing a plurality of logic blocks included in a target program into a plurality of service logic block sets according to a service to which a logic block belongs; concatenating the plurality of the service logic block sets into a service logic chain which is used for executing respective service logic block set in sequence according to an order of the respective service logic block in the service logic chain. Through the method of the present disclosure, service processing logics of different services can be decoupled, which is convenient for distinguishing and maintaining the processing logics of different services and avoiding the mutual interference of the processing logics of different services during execution.

Abstract: The present application discloses an ocular implant delivery device comprising a housing, a rotating wheel, a push needle linkage assembly, a retracting linkage assembly, a sleeve, and a puncture needle. A through hole is provided at a distal end of the housing, and the sleeve extends from an inside of the housing to an outside of the housing through hole and is fixedly connected to the housing. The puncture needle runs through the sleeve, and an end of the puncture needle facing away from a needle tip of the puncture needle is fixedly connected to the retracting linkage assembly. A push needle of the push needle linkage assembly extends through an inner hole of the puncture needle, the rotating wheel is connected to the housing via a rotating shaft, and an opening corresponding to the rotating wheel is provided on the housing.

Abstract: A system and method of performing tensor operations with a multi-step operation processing system in a memory-efficient manner. The method includes the stages of dividing an N-dimensional tensor into a set of tensor slices. The tensor slices consist of one or more consecutive rows. The tensor slices may further be segmented. The tensor slice segments, along with the dependency data, form based on the tensor dependencies are used for an tensor operation computation to generate a first result. Each processed slice segment is fused into a result slice by removing extra data used in the computation. This process is repeated for each slice to be processed and combined into a final processed tensor result.

Abstract: Artificial intelligence is an increasingly important sector of the computer industry. However, artificial intelligence is an extremely computationally intensive field such that performing artificial intelligence calculations can be expensive, time consuming, and energy consuming. Fortunately, many of the calculations required for artificial intelligence applications can be performed in parallel such that specialized linear algebra matrix processors can greatly increase computational performance. But even with linear algebra matrix processors; performance can be limited due to complex data dependencies. Without proper coordination, linear algebra matrix processors may end up idle or spending large amounts of time moving data around. Thus, this document discloses methods for efficiently scheduling linear algebra matrix processors.

Abstract: Artificial intelligence is an increasingly important sector of the computer industry. However, artificial intelligence is extremely computationally intensive field such that it can be expensive, time consuming, and energy consuming. Fortunately, many of the calculations required for artificial intelligence can be performed in parallel such that specialized processors can great increase computational performance. Specifically, artificial intelligence generally requires large numbers of matrix operations to implement neural networks such that specialized Matrix Processor circuits can improve performance. But a neural network is more than a collection of matrix operations; it is a set of specifically coordinated matrix operations with complex data dependencies. Without proper coordination, Matrix Processor circuits may end up idle or spending large amounts of time loading in different weight matrix data.

Abstract: Disclosed are systems and methods for processing a multilayer neural network incorporating skip connections while reducing the memory footprint and processing time of processing a neural network. The method comprises loading within a memory partition with a portion of an input tensor and a portion of layer weights associated with computing a portion of the one or more intermediate layer tensors associated with a first portion of the skip connection tensor. Next, a neural processing unit is used to recompute portions of the skip connection tensor using the portion of the input tensor and associated weights. Upon completion, the memory utilized for the recomputing is free for further computations.

Abstract: Machine learning model scalability with distributed multi-layer processing is disclosed herein. A method for processing and deploying machine learning models that enhances scalability and efficiency by executing a subset of a neural network on each of a plurality of interconnected processing units. The method involves partitioning compute tasks across these processing units to reduce latency, including broadcast and reduction processes for inputs and outputs. It also includes managing the allocation of samples in a batch to specific master processing units within the distributed arrangement and synchronizing computation between fully connected layers within each processing unit. Additionally, the method implements data reduction during the transfer of data across the processing units, wherein data is accumulated with a current processing unit's partial sum as it is transferred to the destination processing unit.

Abstract: Image data of contrast dye in a vessel in a body is acquired for determining flow rate. Based on images acquired under and angle relative to one another, a three-dimensional model of the vessel is constructed and length of a vessel section is determined. A series of at least two images, apart in time, under a first angle is assessed for determining progress of a front of the dye bolus in the vessel in time. In the images, the vessel may be segmented and brightness or a derivative thereof over at least one of time and distance may be assessed to determine the front. Progress distance is mapped to the three-dimensional model, for example by mapping segments from the image to the model, to obtain a more accurate and natural distance of progress over time. Flow rate is determined by natural progress distance over progress time.

Abstract: A formulation, comprising erythropoietin, a buffer ingredient, a surfactant and an osmotic pressure regulator. The present invention also relates to a method for preparing the formulation and the use of the formulation. The formulation is safe and stable and has a long validity period, and the preparation method is simple and easy to implement.

Abstract: Aspects of the present application provide methods and device for use in User Equipment (UE) cooperation. UE cooperation may include the cooperating UEs (CUEs) forwarding traffic to or from one or more target UEs (TUEs) with redundant signal transmissions or receptions. Methods involve a base station transmitting configuration information to at least one cooperative user equipment (CUE) and to a target user equipment (TUE). The configuration information includes an indication of resources for a sidelink (SL) transmission and a redundancy parameter for the SL transmission. The SL transmission is a transmission for the at least one CUE to forward a packet intended for the TUE. The base station also transmits the packet intended for the TUE, to a plurality of UEs comprising the at least one CUE and the TUE.

Abstract: A frame sealing adhesive of the liquid-crystal phase shifter is disposed between two transparent substrates, the frame sealing adhesive encloses a first cavity, a first part of the metal-trace layer is located inside the first cavity, and a second part of the metal-trace layer is located outside the first cavity. The second part is disposed on first surfaces or second surfaces of the two transparent substrates. If the second part is disposed on the first surfaces of the two transparent substrates, metal cushion layers are provided between the frame sealing adhesive and the first surfaces of the two transparent substrates. If the second part is disposed on the second surfaces of the two transparent substrates, the first part and the second part are electrically connected by metal via holes provided in the transparent substrates, and the frame sealing adhesive contacts the first surfaces of the two transparent substrates.

Abstract: Managing thermal capabilities of an information handling system, including identifying a maximum supported ambient temperature of the information handling system; identifying a current ambient temperature of the information handling system; calculating a temperature delta based on the maximum supported ambient temperature of the information handling system and the current ambient temperature of the information handling system; adjusting, based on the temperature delta, one or more thermal control trigger points; adjusting, based on the adjusted thermal control trigger points, a fan speed of a fan of the information handling system; and determining, based on the adjusted fan speed of the fan, an updated cooling capacity associated with the information handling system.

Abstract: Disclosed are semiconductor devices that implement relaxed clock forwarding between logic blocks. In one embodiment the system includes a set of logic blocks forming a first processing path. Another set of logic blocks form additional processing paths. A clock is configured to forward the data between the logic blocks asynchronously in the first processing path in which the data is asynchronously forwarded between logic blocks. This forwarding is asynchronous with the additional processing paths data and clock. The ends or last logic block in each path can be synchronized using a synchronizer component. The synchronizer can be a plurality of asynchronous FIFOs. In one embodiment, logic blocks form a matrix and the processing paths are along the rows or columns of the matrix.

Abstract: Artificial intelligence is an increasingly important sector of the computer industry. One of the most important applications for artificial intelligence is object recognition and classification from digital images. Convolutional neural networks have proven to be a very effective tool for object recognition and classification from digital images. However, convolutional neural networks are extremely computationally intensive thus requiring high-performance processors, significant computation time, and significant energy consumption. To reduce the computation time and energy consumption a “cone of dependency” and “cone of influence” processing techniques are disclosed. These two techniques arrange the computations required in a manner that minimizes memory accesses such that computations may be performed in local cache memory. These techniques significantly reduce the time to perform the computations and the energy consumed by the hardware implementing a convolutional neural network.

Abstract: Artificial intelligence is an increasingly important sector of the computer industry. However, artificial intelligence is extremely computationally intensive field such that it can be expensive, time consuming, and energy consuming. Fortunately, many of the calculations required for artificial intelligence can be performed in parallel such that specialized processors can great increase computational performance. Specifically, artificial intelligence generally requires large numbers of matrix operations to implement neural networks such that specialized matrix processor circuits can improve performance. Thus, this document discloses apparatus and methods for efficiently processing matrix operations.