Boe Technology Group Co., Ltd.

Beijing Boe Optoelectronics Technology Co., Ltd.

Huawei Technologies Co., Limited

Chinese Academy of Sciences

International Business Machines Corporation

Tsinghua University

Juniper Networks, Inc.

Samsung Electronics Co., Ltd.

Hewlett-Packard Development Company, L.p.

Beijing Xiaomi Mobile Software Co., Ltd.

Lenovo (Beijing) Limited

Other

Gm Global Technolgoy Operations, Inc.

United Technologies Corporation

Abb Schweiz Ag

Lei Wang

Xiaochuan Chen

Haisheng Wang

Shengji Yang

Xue Dong

Yingming Liu

Pengcheng Lu

Xiaoliang Ding

Changfeng Li

Rui Li Xu

Li Xiao

Dongni Liu

Ming Yang

Minghua Xuan

Jian Gao

Qian Wang

Weijie Zhao

Wenqing Zhao

Han Yue

Wei Liu

Hongjuan Liu

Hailin Xue

Liang Chen

Pengpeng Wang

Yue Li

Jing Cheng

Ping Zhang

Jing Lv

Xueyou Cao

Xuejuan Gao

Lei Zhang

Fan Yang

Peng Liu

Guangcai Yuan

Yanling Han

Feng Wu

Can Wang

Fang-Chen Cheng

Can Zhang

Jiantao Liu

Yuzhen Guo

Jing Yu

Fuqiang Li

Xueming Pan

Qian Zhang

Lijun Zhao

Wei Zhao

Hao Chen

Hao Wang

Qi Liu

Xiandong Meng

Chuncheng Che

Shaohui Sun

Wei Wang

Hui Jin

Xi Chen

Feng Li

Weiwei Wang

Zhi Hu Wang

Xiao Ling Chen

Jing Liu

Deshan Miao

Fei Wang

Jian Wang

Hong Wang

Wei Li

Fang Lu

Lixiang Xu

Jian Zhang

Hui Li

Wei Zhang

Yu Wang

Li Wang

Shuo Li

Lin Sun

Tao Wang

Dong Chen

Qiubin Gao

Peng Hui Jiang

Jian Qiao Sun

Ying Wang

Qiang Wang

Zhanfeng Cao

Jifeng Tan

Jing Li

Dong Wang

Wei Bai

Runhua Chen

Zheng Liu

Shi Shu

Jie Zhang

Zifeng Wang

Yao Zhang

Wei Ping Li

Peng Wang

Renwei Guo

Xueguang Hao

Xiang-Ming He

Jian-Jun Li

Modem deep neural network (DNN) models have many layers with a single layer potentially involving large matrix multiplications. Such heavy calculation brings challenges to deploy such DNN models on a single edge device, which has relatively limited computation resources. Therefore, multiple and even heterogeneous edge devices may be required for applications with stringent latency requirements. Disclosed in the present patent documents are embodiments of a model scheduling framework that schedules multiple models on a heterogeneous platform. Two different approaches, model first scheduling (MFS) and hardware first scheduling (HFS), are presented to allocate a group of models for a service into corresponding heterogeneous edge devices, including CPU, VPU and GPU. Experimental results prove the effectiveness of the MFS and HFS methods for improving the inference speed of single and multiple AI-based services.

Hardware adaptive multi-model scheduling

An accelerator circuit includes a control interface to receive a stream of instructions, a first memory to store an input data, and an engine circuit. The engine circuit includes a dispatch circuit to decode an instruction of the stream of instructions into a plurality of commands and a plurality of queue circuits. Each of the plurality of queue circuits supports a queue data structure to store a respective one of the plurality of commands decoded from the instruction, and a plurality of command execution circuits. Each of the plurality of command execution circuits is to receive and execute a command extracted from a corresponding one of the plurality of queues.

System and architecture of pure functional neural network accelerator

A display panel includes a display substrate. The display substrate includes: a base substrate; a display structure layer including a plurality of pixel circuits and a plurality of light-emitting devices on the base substrate; and a light detection layer located at a non-light exit side of the light-emitting devices. A pixel circuit in the plurality of pixel circuits is electrically connected to a respective light-emitting device to drive the light-emitting device to emit light. The light detection layer is configured to detect a luminous intensity of at least one light-emitting device. The display substrate further includes: a transparent cover plate at a light exit side of the light-emitting devices; and a polarizer and a first quarter-wave plate between the display substrate and the transparent cover plate. The polarizer is closer to the display substrate than the first quarter-wave plate.

Display panel having light-emitting devices and light detection layer and manufacturing method therefor, and display apparatus

A texture image acquiring device, a display device and a collimator are disclosed. The texture image acquiring device includes a collimator and an image sensor. The collimator includes a lens array and diaphragm component. The diaphragm component includes a first diaphragm layer and a second diaphragm layer. The lens array is configured to allow light rays to be converged and incident on the diaphragm component. The diaphragm component is configured to allow light rays to pass through, and to restrict an angle of light rays capable of passing through the diaphragm component. The image sensor is configured to sense light rays incident on the image sensor for acquiring a texture image.

Texture image acquiring device, display device, and collimator

A system determines a timing matrix corresponding to inference times taken for a number of machine learning (ML) models to be executed by a number of processing resources of a computing device. The processing resources includes at least a first and a second type of processing resources. The system applies a service-specific model-first scheduling scheme or a service-specific hardware-first scheduling scheme to obtain corresponding service-specific mappings. The system determines a best mapping from the corresponding service-specific mappings. The system schedules each of the ML models to a corresponding processing resource from the processing resources according to the best mapping. The system executes the ML models using corresponding mapped processing resources.

Scheduling ML services and models with heterogeneous resources

Disclosed in the embodiments of the present disclosure are a display control method and apparatus, and a device and a storage medium. The method includes: plays a first video in a first display area in a video stream playing interface of a preset application program. The method further receives a switching instruction. When a preset reminding condition is met, the method switches to play a second video in the first display area. The display area for the second video is changed from the first display area to a second display area during the playing process, and the size of the second display area is smaller than the size of the first display area. In addition, the method displays preset prompt information outside the second display area in the video stream playing interface.

Display control method and apparatus, and device and storage medium

A display panel includes a sub-pixel array and a plurality of photosensitive units. The sub-pixel array includes a first sub-pixel, a second sub-pixel and a third sub-pixel that are capable of emitting light of different colors. The plurality of photosensitive units are disposed under gaps of the sub-pixel array. Each of the plurality of photosensitive units includes a photosensitive device, and the photosensitive device includes a photosensitive layer; and an orthographic projection of the photosensitive layer in the photosensitive device on a panel surface of the display panel has overlapping regions with orthographic projections of the first sub-pixel, the second sub-pixel and the third sub-pixel on the panel surface of the display panel.

Display panel and display apparatus

An information transmission processing method, an information transmission processing device and a UE are provided. The information transmission processing method includes: obtaining first DCI and first information, the first DCI including a first NDI; determining that the first NDI has been toggled in the case that the first information meets a first condition; and determining the first NDI to have not been toggled in the case that the first information meets a second condition. The first information includes at least one of: a first scrambling RNTI type of the first DCI; a second scrambling RNTI type of second DCI; a value of the first NDI in the first DCI; or a value of a second NDI in the second DCI.

Information transmission processing method and device, and user equipment

An optical detection apparatus and method for detecting a display panel is provided. The apparatus includes photosensitive units arranged in an array. A minimum distance between adjacent photosensitive units is less than a sub-pixel size of the display panel; the photosensitive units in adjacent rows have a position offset in a first direction; and an offset distance corresponding to the position offset is less than the sub-pixel size of the display panel.

Optical detection apparatus for detecting display panel and method for detecting display panel

Disclosed are a beam determination apparatus, a UE and a network device. A beam determination method performed by the UE includes: determining, by the UE, whether a common beam state is a valid state; and in case that the common beam state is the valid state, determining, by the UE, that at least two channels and/or reference signals of the UE use the same beam for transmission.

Beam determination apparatus, user equipment and network device

Provided is a dynamic latch, including: a substrate; a data transmission unit, including a first transmission gate and a second transmission gate, in which an input end of the first transmission gate is connected to an input end of the second transmission gate, and an output end of the first transmission gate is connected to an output end of the second transmission gate; a data output unit, including a first inverter, in which an input end of the first inverter is connected to the output ends of the two transmission gates, a first region of the substrate is adjacent to a second region thereof and oxide diffusion regions in two regions are continuous, drain regions of the two transmission gates are respectively located on opposite sides in the first region, and source regions of the two transmission gates are located between the drain regions of the two transmission gates.

Dynamic latch, semiconductor chip, computing power board and computing device

Embodiments of the present application provide a method for determining a physical uplink control channel (PUCCH) repetition factor, and a user equipment (UE) and base station, comprising: determining a PUCCH repetition factor according to received target information; said target information comprises first information or second information, said first information comprises a PUCCH resource indicator (PRI) field in downlink control information (DCI), said second information comprises control information carried by the DCI, a media access control control element (MAC CE), or a group common physical downlink control channel (PDCCH). The embodiments of the present application enable a dynamic indication process of the PUCCH repetition factor.

Method for determining PUCCH repetition factor, and UE and base station

The present disclosure discloses a mud pressure wave signal modulation apparatus and method, which relate to the technical field of petroleum drilling engineering. The apparatus includes: a motor (), a valve plate (), a main control circuit (), a motor drive control circuit, and a downhole communication circuit; the valve plate () is driven by the motor () to be rotated around an axial direction perpendicular to a flow direction of drilling fluid; the motor () and the valve plate () are disposed in a drilling fluid channel (), and the drilling fluid channel (), the motor (), the valve plate (), the main control circuit (), the motor drive control circuit and the downhole communication circuit are disposed in a drill collar.

Mud pressure wave signal modulation apparatus and method

The present invention relates to a composite treatment and recovery technique of polluted water body and soil. Iron-rich straw biomass, after being crushed, is mixed and granulated with sludge, and is pyrolytic charred by programmed heating, to obtain bulk loaded zero-valent iron biochar and sludge biochar composite particles, which are packed as fillers in a filled bed or as filters in filter cells of a fixed bed, for effective recovery of complex polluted water, polluted by heavy metals, organics, nitrogen, phosphorus, and the like.

Composite treatment and remediation technology for polluted water and soil

A data marking method, apparatus, system, device, and storage medium are disclosed. A method can include: acquiring multiple pictures whose contents are continuous, wherein the multiple pictures contain at least one same object; for each object, determining a position offset of the object by using position information of the object in two adjacent pictures, wherein the two adjacent pictures include a first previous picture and a second previous picture, the second previous picture is a picture before and adjacent to a picture to be marked in time sequence; the first previous picture is a picture before and adjacent to the second previous picture in time sequence; determining estimated position information of the object in the picture to be marked based on the position information of the second previous picture and the position offset; marking the object in the picture to be marked based on the estimated position information.

Data marking method, apparatus, system, device and storage medium

Growing community of inventors

Beijing, China

Lei Wang