Patents by Inventor Junpeng WANG

Junpeng WANG has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Patent number: 11861324
    Abstract: Provided is a method for normalizing embeddings for cross-embedding alignment. The method may include applying mean centering to the at least one embedding set, applying spectral normalization to the at least one embedding set, and/or applying length normalization to the at least one embedding set. Spectral normalization may include decomposing the at least one embedding set, determining an average singular value of the at least one embedding set, determining a respective substitute singular value for each respective singular value of a diagonal matrix, and/or replacing the at least one embedding set with a product of the at least one embedding set, a right singular vector, and an inverse of the substitute diagonal matrix. The mean centering, spectral normalization, and/or length normalization may be iteratively repeated for a configurable number of iterations. A system and computer program product are also disclosed.
    Type: Grant
    Filed: May 25, 2022
    Date of Patent: January 2, 2024
    Assignee: Visa International Service Association
    Inventors: Yan Zheng, Michael Yeh, Junpeng Wang, Wei Zhang, Liang Wang, Hao Yang, Prince Osei Aboagye
  • Publication number: 20230365743
    Abstract: A monomer capable forming a polymer through ring-opening metathesis polymerization and capable of depolymerization thereafter through ring-closing metathesis, wherein the monomer comprises a cycloalkene having a fused ring attached thereto which decreases the ring strain energy to 5.3 kcal/mol or lower as compared to the same cycloalkene without a fused ring having a ring strain energy above 5.3 kcal/mol.
    Type: Application
    Filed: September 13, 2021
    Publication date: November 16, 2023
    Inventors: Junpeng Wang, Devavrat Sathe
  • Publication number: 20230351215
    Abstract: A method includes extracting, by an analysis computer, a plurality of first datasets from a plurality of graph snapshots using a graph structural learning module. The analysis computer can then extract a plurality of second datasets from the plurality of first datasets using a temporal convolution module across the plurality of graph snapshots.
    Type: Application
    Filed: September 17, 2021
    Publication date: November 2, 2023
    Applicant: VISA INTERNATIONAL SERVICE ASSOCIATION
    Inventors: Jiarui Sun, Mengting Gu, Junpeng Wang, Yanhong Wu, Liang Wang, Wei Zhang
  • Publication number: 20230340174
    Abstract: A macromonomer capable of forming a graft polymer through graft-through polymerization, the macromonomer including a plurality of monomer units including a cycloalkene having a fused ring attached thereto to form a cycloalkene-fused ring monomer where the fused ring decreases the ring strain energy of the cycloalkene to a lower ring strain energy state of 5.3 kcal/mol or lower and where the cycloalkene of the cycloalkene-fused ring monomer is capable of isomerization into a higher ring strain energy state before graft-through polymerization, and at least one polymer sidechain is bonded to the fused ring. Graft polymers formed by the macromonomer include graft polymers and graft copolymers including statistical and block copolymers. The graft (co)polymers are capable of depolymerization under mild conditions into reusable monomer units.
    Type: Application
    Filed: January 16, 2023
    Publication date: October 26, 2023
    Inventors: Junpeng Wang, Zeyu Wang, Seiyoung Yoon
  • Publication number: 20230274135
    Abstract: A method, system, and computer program product is provided for embedding compression and reconstruction. The method includes receiving embedding vector data comprising a plurality of embedding vectors. A beta-variational autoencoder is trained based on the embedding vector data and a loss equation. The method includes determining a respective entropy of a respective mean and a respective variance of each respective dimension of a plurality of dimensions. A first subset of the plurality of dimensions is determined based on the respective entropy of the respective mean and the respective variance for each respective dimension of the plurality of dimensions. A second subset of the plurality of dimensions is discarded based on the respective entropy of the respective mean and the respective variance for each respective dimension of the plurality of dimensions. The method includes generating a compressed representation of the embedding vector data based on the first subset of dimensions.
    Type: Application
    Filed: October 21, 2022
    Publication date: August 31, 2023
    Inventors: Haoyu Li, Junpeng Wang, Liang Wang, Yan Zheng, Wei Zhang
  • Publication number: 20230214177
    Abstract: Provided is a method for normalizing embeddings for cross-embedding alignment. The method may include applying mean centering to the at least one embedding set, applying spectral normalization to the at least one embedding set, and/or applying length normalization to the at least one embedding set. Spectral normalization may include decomposing the at least one embedding set, determining an average singular value of the at least one embedding set, determining a respective substitute singular value for each respective singular value of a diagonal matrix, and/or replacing the at least one embedding set with a product of the at least one embedding set, a right singular vector, and an inverse of the substitute diagonal matrix. The mean centering, spectral normalization, and/or length normalization may be iteratively repeated for a configurable number of iterations. A system and computer program product are also disclosed.
    Type: Application
    Filed: May 25, 2022
    Publication date: July 6, 2023
    Inventors: Yan Zheng, Michael Yeh, Junpeng Wang, Wei Zhang, Liang Wang, Hao Yang, Prince Osei Aboagye
  • Patent number: 11694064
    Abstract: A method for local approximation of a predictive model may include receiving unclassified data associated with a plurality of unclassified data items. The unclassified data may be classified based on a first predictive model to generate classified data. A first data item may be selected from the classified data. A plurality of generated data items associated with the first data item may be generated using a generative model. The plurality of generated data items may be classified based on the first predictive model to generate classified generated data. A second predictive model may be trained with the classified generated data. A system and computer program product are also disclosed.
    Type: Grant
    Filed: September 27, 2022
    Date of Patent: July 4, 2023
    Assignee: Visa International Service Association
    Inventors: Liang Gou, Junpeng Wang, Wei Zhang, Hao Yang
  • Publication number: 20230186078
    Abstract: A method for evaluating a RNN-based deep learning model includes: receiving model data generated by the RNN-based model, the model data including a plurality of events associated with a plurality of states; generating a first GUI based on the events and states including a chart visually representing a timeline for the events in relation to a parameter value; generating a second GUI including a point chart visually representing a two-dimensional projection of the multi-dimensional intermediate data, each point of the point chart representing a time step and an event from the time step, based on multi-dimensional intermediate data between transformations in the model that connect a state to an event; and perturbing the environment at a time step based on user interaction with at least one of the first and second GUIs.
    Type: Application
    Filed: April 30, 2021
    Publication date: June 15, 2023
    Inventors: Junpeng Wang, Wei Zhang, Hao Yang, Michael Yeh, Liang Wang
  • Publication number: 20230167211
    Abstract: A monomer capable forming a polymer through ring-opening metathesis polymerization and capable of depolymerization thereafter through ring-closing metathesis, wherein the monomer comprises a cycloalkene having a fused ring attached thereto to form a cycloalkene-fused ring monomer, wherein the fused ring decreases the ring strain energy of the cycloalkene to a lower ring strain energy state of 5.3 kcal/mol or lower as compared to the same cycloalkene without the fused ring having a ring strain energy above 5.3 kcal/mol and wherein the cycloalkene-fused ring monomer is capable of isomerization into a higher ring strain energy state before polymerization and method of synthesizing the monomer. The monomer is further capable of forming chemically recyclable to monomers block copolymers.
    Type: Application
    Filed: September 30, 2022
    Publication date: June 1, 2023
    Inventors: Junpeng Wang, Devavrate Sathe, Junfeng Zhou, Hanlin Chen
  • Publication number: 20220414662
    Abstract: A method for detecting collusive transaction fraud includes: generating a merchant baseline including a transaction data baseline and a time series baseline; extracting time series data of the first merchant system; generating a first score and second score with a deep learning model; generating a first merchant risk score of the first merchant system based on the first and second scores; in response to determining that the first merchant risk score satisfies the threshold, determining a plurality of related entities related to the first merchant system; and classifying the first merchant system and at least one related entity of the plurality of related entities in a first group risk class based on at least one risk score of the at least one related entity.
    Type: Application
    Filed: June 25, 2021
    Publication date: December 29, 2022
    Inventors: Shi Cao, Chiranjeet Chetia, Liang Wang, Junpeng Wang, Morvarid Jamalian
  • Patent number: 11487997
    Abstract: A method for local approximation of a predictive model may include receiving unclassified data associated with a plurality of unclassified data items. The unclassified data may be classified based on a first predictive model to generate classified data. A first data item may be selected from the classified data. A plurality of generated data items associated with the first data item may be generated using a generative model. The plurality of generated data items may be classified based on the first predictive model to generate classified generated data. A second predictive model may be trained with the classified generated data. A system and computer program product are also disclosed.
    Type: Grant
    Filed: October 4, 2019
    Date of Patent: November 1, 2022
    Assignee: Visa International Service Association
    Inventors: Liang Gou, Junpeng Wang, Wei Zhang, Hao Yang
  • Publication number: 20220201548
    Abstract: A method for optimizing mobile network congestion includes detecting that the UE is in a congestion mode; determining whether the UE enters the RRC connection mode in a first priority cell; determining whether shortening at least one timer of the UE is able to improve the mobile network congestion if the UE does not enter the RRC connection mode; determining whether reporting the larger data size of the BSR to the mobile network is able to improve the mobile network congestion if the UE enters the RRC connection mode; reselecting a second priority cell if shortening the at least one timer of the UE is not able to improve the mobile network congestion; and handing over to the second priority cell if reporting the larger data size of the BSR to the mobile network is not able to improve the mobile network congestion.
    Type: Application
    Filed: November 30, 2021
    Publication date: June 23, 2022
    Inventor: Junpeng WANG
  • Publication number: 20210224648
    Abstract: Described are a system, method, and computer program product for multivariate event prediction using multi-stream recurrent neural networks. The method includes receiving event data from a sample time period and generating feature vectors for each subperiod of each day. The method also includes providing the feature vectors as inputs to a set of first recurrent neural network (RNN) models and generating first outputs for each RNN node. The method further includes merging the first outputs for each same subperiod to form aggregated time-series layers. The method further includes providing the aggregated time-series layers as an input to a second RNN model and generating final outputs for each RNN node of the second RNN model.
    Type: Application
    Filed: January 14, 2021
    Publication date: July 22, 2021
    Inventors: Zhongfang Zhuang, Michael Yeh, Liang Wang, Wei Zhang, Junpeng Wang
  • Publication number: 20210209604
    Abstract: Provided is a method for detecting group activities in a network. The method may include receiving interaction data associated with a plurality of interactions. For each account identifier associated with at least one interaction, a value may be determined for each of a first set of categories, and a vector may be generated based on the value for each category. The length of each vector may be determined. At least one relational graph may be generated based on the interaction data. Each relational graph may be associated with a respective category of a second set of categories. At least one cluster of nodes may be determined based on the relational graph(s). A score for each cluster may be determined based on the length of the vector associated with the account identifier of each node of the cluster of nodes. A system and computer program product are also disclosed.
    Type: Application
    Filed: December 30, 2020
    Publication date: July 8, 2021
    Inventors: Liang Wang, Junpeng Wang, Chiranjeet Chetia, Shi Cao, Harishkumar Sundarji Majithiya, Roshni Ann Samuel, Minghua Xu, Wei Zhang, Hao Yang
  • Publication number: 20210091447
    Abstract: A microwave circulator/isolator includes PINs, a shell, and a locating dielectric slice, wherein a laminated assembly is arranged in the shell, notches are formed in the peripheral wall of the shell, and the laminated assembly includes a central conductor electrically connected to the PINs which are fixed to the locating dielectric slice arranged in the shell. The microwave circulator/isolator simplifies the production process, lowers the machining difficulty of the shell, reduces the production cost, and improves the production efficiency.
    Type: Application
    Filed: March 18, 2019
    Publication date: March 25, 2021
    Inventors: Fei XIONG, Wei ZHANG, Huanhuan GUO, Zongwei TANG, Jing NI, Junpeng WANG, Yuanyong XU
  • Publication number: 20210083698
    Abstract: Example signal processing apparatus and signal processing methods are described. One example signal processing apparatus includes a signal matching circuit, a first signal processing branch, and a second signal processing branch. An output end of the signal matching circuit is separately coupled to an input end of the first signal processing branch and an input end of the second signal processing branch, and an output end of the first signal processing branch is coupled to the input end of the second signal processing branch.
    Type: Application
    Filed: November 25, 2020
    Publication date: March 18, 2021
    Inventors: Junpeng WANG, Tengteng JI, Yugui XUE, Caixia CUI
  • Patent number: 10946459
    Abstract: A magnetorheological support method for blisk processing is disclosed. In the method, a fork structure and a soft film are used to wrap magnetorheological fluid. The magnetorheological fluid is used for flow filling under certain pressure. The bulged soft film can conduct shape matching on the surface of a blisk blade. The magnetorheological fluid can be cured through magnetic field excitation, thereby ensuring the flexible support for a weak rigid component. Electric permanent magnets are symmetrically arranged at both ends of the fork structure to construct a uniform magnetic field that can realize a global excitation of magnetorheological fluid, so that the magnetorheological fluid works in a shear mode to achieve damping force controlling by magnetic field. The solid-liquid conversion of the magnetorheological fluid is controlled by an electric permanent magnet field.
    Type: Grant
    Filed: September 29, 2019
    Date of Patent: March 16, 2021
    Assignee: DALIAN UNIVERSITY OF TECHNOLOGY
    Inventors: Haibo Liu, Qi Luo, Junpeng Wang, Te Li, Kuo Liu, Yongqing Wang
  • Publication number: 20200376571
    Abstract: A magnetorheological support method for blisk processing is disclosed. In the method, a fork structure and a soft film are used to wrap magnetorheological fluid. The magnetorheological fluid is used for flow filling under certain pressure. The bulged soft film can conduct shape matching on the surface of a blisk blade. The magnetorheological fluid can be cured through magnetic field excitation, thereby ensuring the flexible support for a weak rigid component. Electric permanent magnets are symmetrically arranged at both ends of the fork structure to construct a uniform magnetic field that can realize a global excitation of magnetorheological fluid, so that the magnetorheological fluid works in a shear mode to achieve damping force controlling by magnetic field. The solid-liquid conversion of the magnetorheological fluid is controlled by an electric permanent magnet field.
    Type: Application
    Filed: September 29, 2019
    Publication date: December 3, 2020
    Inventors: Haibo LIU, Qi LUO, Junpeng WANG, Te LI, Kuo LIU, Yongqing WANG
  • Publication number: 20200110982
    Abstract: A method for local approximation of a predictive model may include receiving unclassified data associated with a plurality of unclassified data items. The unclassified data may be classified based on a first predictive model to generate classified data. A first data item may be selected from the classified data. A plurality of generated data items associated with the first data item may be generated using a generative model. The plurality of generated data items may be classified based on the first predictive model to generate classified generated data. A second predictive model may be trained with the classified generated data. A system and computer program product are also disclosed.
    Type: Application
    Filed: October 4, 2019
    Publication date: April 9, 2020
    Inventors: Liang Gou, Junpeng Wang, Wei Zhang, Hao Yang
  • Patent number: 10615827
    Abstract: The present disclosure discloses a local oscillator feedthrough signal correction apparatus, including a microprocessor control unit, a first digital-to-analog converter, a second digital-to-analog converter, a mixer, a local oscillator, a signal output line, a signal splitter, and a detector tube. The signal splitter is disposed in the signal output line, and the first digital-to-analog converter and the second digital-to-analog converter are configured to provide the mixer with quadrature direct current components VI and VQ used for local oscillator feedthrough signal correction. The mixer outputs a local oscillator feedthrough signal to the signal output line. The signal splitter obtains the local oscillator feedthrough signal by means of splitting, and the detector tube detects the local oscillator feedthrough signal.
    Type: Grant
    Filed: May 10, 2018
    Date of Patent: April 7, 2020
    Assignee: HUAWEI TECHNOLOGIES CO., LTD.
    Inventors: Tengteng Ji, Junpeng Wang, Bin Ding, Haiming Fan