公开(公告)号：US20230009961A1

公开(公告)日：2023-01-12

申请号：US17806614

申请日：2022-06-13

Applicant: Huazhong University of Science and Technology

Inventor： Jiang XIAO ,  Feng CHENG ,  Junpei NI ,  Wenhui YANG ,  Hai JIN

IPC: H04L9/32 ,  H04L9/00 ,  H04L9/06

Abstract: The present invention relates to a graphic-blockchain-orientated sharding storage apparatus, at least comprising a first sharding module and a second sharding module, wherein the first sharding module shards nodes having different resource capacity levels based on ledger data organized using a DAG structure, and the second sharding module assigns transactions to the shards matching with execution difficulty levels of the transactions, so that each said transaction is processed and stored in the shard corresponding thereto. The present invention incorporates the sharding technology into a graphic blockchain to provide a graphic-blockchain-orientated sharding storage method, so as to reduce pressure in terms of data storage and transaction processing on nodes of the graphic blockchain system. In addition, nodes, transactions, and data are dynamically divided according to resource heterogeneity among nodes, so as to further enhance performance of the graphic blockchain system while achieving efficient use of resources.

公开(公告)号：US20220318688A1

公开(公告)日：2022-10-06

申请号：US17644425

申请日：2021-12-15

Applicant: Huazhong University of Science and Technology

Inventor： Jiang XIAO ,  Xiaohai DAI ,  Huichuwu LI ,  Chen YU ,  Hai JIN

IPC: G06N20/20

Abstract: The present invention relates a method and a system for cross-chain consensus oriented to federated learning, comprising: conducting intra-cluster single-chain federated learning and collecting local update information; sending updates after consensus to a second federation so as to execute cross-cluster gradient exchange; receiving a verification result of cross-cluster gradient update consensus fed back from the second federation; and conducting local model update based on the verification result. After implementation of the update consensus, the present invention provides rewards and punishments based on the contributions of the cluster representatives, thereby encouraging the cluster representatives in the computing nodes to act honestly, so that the participants can actively help the model update.

公开(公告)号：US20220012155A1

公开(公告)日：2022-01-13

申请号：US17247237

申请日：2020-12-04

Applicant: HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY

Inventor： Jiang XIAO ,  Huichuwu LI ,  Minrui Wu ,  Hai JIN

IPC: G06F11/34 ,  G06F21/60 ,  G06N20/00

Abstract: The present invention relates to an activity recognition system balanced between versatility and individuation, comprising a communication framework jointly formed by a data collecting terminal, a computing device, and a cloud computing platform, the activity recognition system uses the communication framework to conduct personnel activity recognition and model updating, and the edge computing device further comprises a model training module and an activity recognition module, and the model training module retrieves a local activity recognition model by continuously verifying user IDs, and uses the first data to train a versatile network structure and an individualized network structure of the local activity recognition model in a way that individuation features of the user and versatility features of the model are fused with each other, so that the personnel activity recognition process conducted by the activity recognition module.

公开(公告)号：US20200272907A1

公开(公告)日：2020-08-27

申请号：US16748284

申请日：2020-01-21

Applicant: Huazhong University of Science and Technology

Inventor： Hai JIN ,  Xiaofei LIAO ,  Long ZHENG ,  Haikun LIU ,  Xi GE

IPC: G06N3/10 ,  G06N3/08 ,  G06N3/04 ,  G06F9/50

Abstract: A deep learning heterogeneous computing method based on layer-wide memory allocation, at least comprises steps of: traversing a neural network model so as to acquire a training operational sequence and a number of layers L thereof; calculating a memory room R1 required by data involved in operation at the ith layer of the neural network model under a double-buffer configuration, where 1≤i≤L; altering a layer structure of the ith layer and updating the training operational sequence; distributing all the data across a memory room of the CPU and the memory room of the GPU according to a data placement method; performing iterative computation at each said layer successively based on the training operational sequence so as to complete neural network training.

公开(公告)号：US20190213029A1

公开(公告)日：2019-07-11

申请号：US16135598

申请日：2018-09-19

Applicant: Huazhong University of Science and Technology

Inventor： Fangming LIU ,  Hai JIN ,  Xiaoyao LI

IPC: G06F9/455 ,  G06F9/50 ,  G06F15/78

Abstract: The present invention relates to an FPGA-based method and system for network function accelerating. The method comprises: building a network function accelerating system that includes a physical machine and an accelerator card connected through a PCIe channel, wherein the physical machine includes a processor and the accelerator card includes an FPGA, in which the accelerator card serves to provide network function accelerating for the processor; the processor being configured to: when it requires the accelerator card to provide network function accelerating, check whether there is any required accelerator module present in the FPGA, and if yes, acquire an accelerating function ID corresponding to the required accelerator module, and if not, select at least one partial reconfigurable region in the FPGA and configure it into the required accelerator module and generate a corresponding accelerating function ID; and/or sending an accelerating request to the FPGA.

公开(公告)号：US20190207763A1

公开(公告)日：2019-07-04

申请号：US16169377

申请日：2018-10-24

Applicant: Huazhong University of Science and Technology

Inventor： Hai JIN ,  Peng XU ,  Shuanghong HE ,  Deqing ZOU

IPC: H04L9/30 ,  H04L9/06 ,  G06F17/30 ,  G06F7/58

Abstract: The present invention involves with a method of searchable public-key encryption, a system and server using the method.

公开(公告)号：US20190205193A1

公开(公告)日：2019-07-04

申请号：US16135480

申请日：2018-09-19

Applicant: Huazhong University of Science and Technology

Inventor： Song WU ,  Zhuang XIONG ,  Hai JIN

IPC: G06F11/00 ,  G06F11/22 ,  G06F11/34 ,  G06F11/30 ,  G06N99/00 ,  G06K9/62

CPC classification number: G06F11/008 ,  G06F11/2221 ,  G06F11/2268 ,  G06F11/2273 ,  G06F11/3034 ,  G06F11/3485 ,  G06F2201/81 ,  G06K9/6269 ,  G06N20/00

Abstract: An S.M.A.R.T. threshold optimization method used for disk failure detection includes the steps of: analyzing S.M.A.R.T. attributes based on correlation between S.M.A.R.T. attribute information about plural failed and non-failed disks and failure information and sieving out weakly correlated attributes and/or strongly correlated attributes; and setting threshold intervals, multivariate thresholds and/or native thresholds corresponding to the S.M.A.R.T. attributes based on distribution patterns of the strongly or weakly correlated attributes. As compared to reactive fault tolerance, the disclosed method has no negative effects on reading and writing performance of disks and performance of storage systems as a whole. As compared to the known methods that use native disk S.M.A.R.T. thresholds, the disclosed method significantly improves disk failure detection rate with a low false alarm rate. As compared to disk failure forecast based on machine learning algorithm, the disclosed method has good interpretability and allows easy adjustment of its forecast performance.