Patents by Inventor Yuhei UMEDA

Yuhei UMEDA 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: 11438277
    Abstract: An allocation method executed by a computer includes dividing each of a plurality of pieces of time-series data into a plurality of segments, allocating a label to each of the pieces of time-series data based on features of each segment in the pieces of time-series data, and allocating a predetermined segment in time-series data, included in the pieces of time-series data, with a label allocated to the time-series data to which the predetermined segment belongs.
    Type: Grant
    Filed: February 26, 2020
    Date of Patent: September 6, 2022
    Assignees: FUJITSU LIMITED, NATIONAL UNIVERSITY CORPORATION KUMAMOTO UNIVERSITY
    Inventors: Yasushi Sakurai, Yasuko Matsubara, Yasuaki Irifune, Saeru Yamamuro, Kouki Kawabata, Akira Ura, Takashi Katoh, Yuhei Umeda
  • Patent number: 11423298
    Abstract: A determination apparatus extracts a plurality of specific events that have values greater than an event determination threshold from among a plurality of events that have occurred in chronological order. The determination apparatus generates a feature amount related to adjacent occurrence intervals of the plurality of specific events, using the plurality of specific events. The determination apparatus generates array data corresponding to the plurality of events using points each having components of the event determination threshold and the feature amount, while changing the event determination threshold. The determination apparatus determines a type of the plurality of events using the array data.
    Type: Grant
    Filed: November 21, 2018
    Date of Patent: August 23, 2022
    Assignee: FUJITSU LIMITED
    Inventor: Yuhei Umeda
  • Publication number: 20220245405
    Abstract: A deterioration suppression device generates a plurality of trained machine learning models having different characteristics on the basis of each training data included in a first training data set and assigned with a label indicating correct answer information. In a case where estimation accuracy of label estimation with respect to input data to be estimated by any trained machine learning model among the plurality of trained machine learning models becomes lower than a predetermined standard, the deterioration suppression device generates a second training data set including a plurality of pieces of training data using an estimation result by a trained machine learning model with the estimation accuracy equal to or higher than the predetermined standard. The deterioration suppression device executes re-learning of the trained machine learning model with the estimation accuracy lower than the predetermined standard using the second training data set.
    Type: Application
    Filed: April 25, 2022
    Publication date: August 4, 2022
    Applicant: FUJITSU LIMITED
    Inventors: TAKASHI KATOH, Kento UEMURA, Suguru YASUTOMI, Tomohiro Hayase, YUHEI UMEDA
  • Publication number: 20220237407
    Abstract: A non-transitory computer-readable storage medium storing an estimation program that causes a computer to execute a process includes specifying representative points of each of training clusters that corresponds to each of labels targeted for estimation; setting boundaries between each of input clusters under a condition that a number of the input clusters and a number of the representative points coincide with each other, the input clusters being generated by clustering in a feature space for input data; acquiring estimation results for the labels with respect to the input data based on a correspondence relationship between the input clusters and the training clusters based on the boundaries; and estimating determination accuracy for the labels by using the machine learning model with respect to the input data based on the estimation results.
    Type: Application
    Filed: April 19, 2022
    Publication date: July 28, 2022
    Applicant: FUJITSU LIMITED
    Inventors: YUHEI UMEDA, TAKASHI KATOH, Yuichi Ike, Mari Kajitani, Masatoshi Takenouchi
  • Publication number: 20220147834
    Abstract: A non-transitory computer-readable recording medium storing an information processing program for causing a computer to execute processing including: acquiring a plurality of points on a Pareto front of a multi-objective optimization problem; fitting a Bezier simplex defined using a plurality of control points to the plurality of points on the Pareto front; and determining whether the Pareto front is degenerated based on a positional relationship among the plurality of control points in the Bezier simplex after the fitting.
    Type: Application
    Filed: September 2, 2021
    Publication date: May 12, 2022
    Applicant: FUJITSU LIMITED
    Inventors: KEN KOBAYASHI, YUHEI UMEDA
  • Publication number: 20210390623
    Abstract: A non-transitory computer-readable recording medium has stored therein a program that causes a computer to execute a process, the process including determining numerical values indicating features at respective timings having a predetermined time interval with respect to time-series data to be analyzed, numbers of the numerical values at the respective timings being made same, and generating an attractor related to the time-series data based on the determined numerical values.
    Type: Application
    Filed: May 26, 2021
    Publication date: December 16, 2021
    Applicant: FUJITSU LIMITED
    Inventors: Masaru TODORIKI, Yuhei UMEDA, Ken KOBAYASHI, Koji MARUHASHI
  • Patent number: 11120302
    Abstract: A method includes: executing a first generation process that includes generating a Betti number series corresponding to a contribution rate by performing persistent homology processing on a first point set, the first point set being generated by using a plurality of pieces of time series data and the contribution rate of each of the plurality of pieces of time series data, each of points included in the first point set being represented by coordinates; executing a second generation process that includes generating a characteristic image from a plurality of the Betti number series, the plurality of Betti number series being generated by performing the first generation process on each of the plurality of contribution rates; and executing a third generation process that includes generating machine learning data in which the characteristic image and a classification corresponding to the plurality of pieces of time series data are associated with each other.
    Type: Grant
    Filed: August 29, 2019
    Date of Patent: September 14, 2021
    Assignee: FUJITSU LIMITED
    Inventor: Yuhei Umeda
  • Patent number: 11024022
    Abstract: A non-transitory computer-readable recording medium storing a program that causes a computer to execute a procedure, the procedure includes generating, for each of a plurality of wafers, extended coordinates including a position on the wafer and a value calculated from a distance from a center of the wafer and a contribution parameter, for each defect on the wafer by using information of a defect position on the wafer, generating a Betti number group by persistent homology processing for a plurality of extended coordinates generated for each of the plurality of wafers generating, for each of the plurality of wafers, a defect pattern image from a plurality of Betti number groups generated for the plurality of values of contribution parameter, and generating machine learning data associating a plurality of defect pattern images generated for the plurality of wafers with determination information associated with the plurality of wafers.
    Type: Grant
    Filed: April 5, 2019
    Date of Patent: June 1, 2021
    Assignee: FUJITSU LIMITED
    Inventors: Yuhei Umeda, Tsutomu Ishida
  • Patent number: 11023562
    Abstract: A non-transitory computer-readable recording medium stores therein an analysis program that causes a computer to execute a process including: dividing a Betti number sequence into a plurality of Betti number sequences, the Betti number sequence being included in a result of a persistent homology process performed on time series data, the plurality of Betti number sequences corresponding to different dimension of the Betti number sequence; and performing an analysis on each of the plurality of Betti number sequences.
    Type: Grant
    Filed: July 6, 2018
    Date of Patent: June 1, 2021
    Assignee: FUJITSU LIMITED
    Inventors: Ken Kobayashi, Yuhei Umeda, Masaru Todoriki, Hiroya Inakoshi
  • Publication number: 20200397330
    Abstract: A detection device generates plural attractors based on brainwave data. Subsequently, the detection device calculates a Betti number by subjecting the attractors to persistent homology transform. The detection device determines an onset of encephalopathy based on a first order component of a Betti sequence calculated based on the Betti number.
    Type: Application
    Filed: June 11, 2020
    Publication date: December 24, 2020
    Applicant: FUJITSU LIMITED
    Inventors: KEN KOBAYASHI, YUHEI UMEDA, Yoshiaki Ikai, KAZUAKI HIRAOKA, Tomoyuki Tsunoda, Yoshimasa KADOOKA
  • Patent number: 10839258
    Abstract: A detection device adds, with regard to each of a plurality of sets of time-series data including a plurality of items, a time-shift term to at least any of the plurality of items included in each of the plurality of sets of time-series data. The detection device generates a plurality of attractors from the plurality of sets of time-series data to which the time-shift term has been added. The detection device generates a plurality of Betti sequences from each of the plurality of attractors by executing a persistent homology transformation on each of the plurality of attractors, each of the plurality of Betti sequences indicating a correspondence relationship between a Betti number and a scale value has been used for the persistent homology transformation. The detection device detects a state change in the plurality of sets of time-series data based on a time change in the plurality of Betti sequences.
    Type: Grant
    Filed: January 31, 2019
    Date of Patent: November 17, 2020
    Assignee: FUJITSU LIMITED
    Inventors: Masaru Todoriki, Yuhei Umeda, Ken Kobayashi
  • Publication number: 20200311587
    Abstract: A method for extracting features, the method being implemented by a computer, the method includes generating attractors from time series data having a cyclic characteristic; generating a persistence diagram by performing persistent homology conversion for the attractors; changing a degree of influence with respect to individual items of data in the persistence diagram in accordance with a time of existence or an appearance time of a hole generated by performing the persistent homology conversion; and extracting features of the time series data from the changed persistence diagram in which the degree of influence has been changed.
    Type: Application
    Filed: March 24, 2020
    Publication date: October 1, 2020
    Applicant: FUJITSU LIMITED
    Inventors: TOMOYUKI TSUNODA, YUHEI UMEDA
  • Publication number: 20200280525
    Abstract: An allocation method executed by a computer includes dividing each of a plurality of pieces of time-series data into a plurality of segments, allocating a label to each of the pieces of time-series data based on features of each segment in the pieces of time-series data, and allocating a predetermined segment in time-series data, included in the pieces of time-series data, with a label allocated to the time-series data to which the predetermined segment belongs.
    Type: Application
    Filed: February 26, 2020
    Publication date: September 3, 2020
    Inventors: Yasushi SAKURAI, Yasuko MATSUBARA, Yasuaki IRIFUNE, Saeru YAMAMURO, Kouki KAWABATA, Akira URA, TAKASHI KATOH, YUHEI UMEDA
  • Patent number: 10692256
    Abstract: A non-transitory computer-readable recording medium stores therein a visualization program that causes a computer to execute a process including: generating a plurality of conversion vectors, from a plurality of vectors generated from plural pieces of input data, by a dimensional compression in a positional relation between the plurality of vectors; and plotting the plurality of conversion vectors.
    Type: Grant
    Filed: October 30, 2018
    Date of Patent: June 23, 2020
    Assignee: FUJITSU LIMITED
    Inventor: Yuhei Umeda
  • Patent number: 10635975
    Abstract: A disclosed machine learning method includes: calculating a first output error between a label and an output in a case where dropout in which values are replaced with 0 is executed for a last layer of a first channel among plural channels in a parallel neural network; calculating a second output error between the label and an output in a case where the dropout is not executed for the last layer of the first channel; and identifying at least one channel from the plural channels based on a difference between the first output error and the second output error to update parameters of the identified channel.
    Type: Grant
    Filed: December 19, 2016
    Date of Patent: April 28, 2020
    Assignee: FUJITSU LIMITED
    Inventor: Yuhei Umeda
  • Publication number: 20200074281
    Abstract: A learning device performs learning by an autoencoder, using waveform data with changes over time that is obtained from an intrinsic movement of an object. The learning device performs persistent homology conversion to calculate a change in the number of connected component according to a threshold change in a value direction for the waveform data. The learning device determines abnormality based on a determination result of a learner to which output data of the autoencoder obtained from the waveform data and output data obtained from the persistent homology conversion are input, and in which machine learning about abnormality of the waveform data has been performed.
    Type: Application
    Filed: August 29, 2019
    Publication date: March 5, 2020
    Applicant: FUJITSU LIMITED
    Inventors: Meryll Dindin, YUHEI UMEDA, Frederic Chazal
  • Publication number: 20190385020
    Abstract: A method includes: executing a first generation process that includes generating a Betti number series corresponding to a contribution rate by performing persistent homology processing on a first point set, the first point set being generated by using a plurality of pieces of time series data and the contribution rate of each of the plurality of pieces of time series data, each of points included in the first point set being represented by coordinates; executing a second generation process that includes generating a characteristic image from a plurality of the Betti number series, the plurality of Betti number series being generated by performing the first generation process on each of the plurality of contribution rates; and executing a third generation process that includes generating machine learning data in which the characteristic image and a classification corresponding to the plurality of pieces of time series data are associated with each other.
    Type: Application
    Filed: August 29, 2019
    Publication date: December 19, 2019
    Applicant: FUJITSU LIMITED
    Inventor: YUHEI UMEDA
  • Publication number: 20190279039
    Abstract: A non-transitory computer-readable recording medium stores therein a learning program that causes a computer to execute a process including: setting each of scores to each of a plurality of sets of unlabeled data with regard to each of labels used in a plurality of sets of labeled data based on a distance of each of the plurality of sets of unlabeled data with respect to each of the labels; and causing a learning model to learn using a neural network by using the plurality of sets of labeled data respectively corresponding to the labels of the plurality of sets of labeled data, and the plurality of sets of unlabeled data respectively corresponding to the scores of the plurality of sets of unlabeled data with regard to the labels.
    Type: Application
    Filed: February 26, 2019
    Publication date: September 12, 2019
    Applicant: FUJITSU LIMITED
    Inventor: YUHEI UMEDA
  • Publication number: 20190279085
    Abstract: A non-transitory computer-readable recording medium stores therein a learning program that causes a computer to execute a process including: setting a label vector having one or a plurality of labels as components to corresponding data to be learned ; and learning a learning model including a neural network using the data to be learned and the label vector correspondingly set to the data to be learned.
    Type: Application
    Filed: February 27, 2019
    Publication date: September 12, 2019
    Applicant: FUJITSU LIMITED
    Inventor: YUHEI UMEDA
  • Patent number: 10408157
    Abstract: A non-transitory computer-readable recording medium stores a data-acquisition-instruction generating program that causes a computer to execute a process including: first generating a plurality of change curves of each of control parameters based on requisite density information, the requisite density information being related to a data measurement density in a data measurement region specified by a combination of a plurality of control parameters, the plurality of control parameters being used by a device subject to the data measurement; and second generating a data acquisition instruction to perform measurement at a plurality of measurement points with respect to the device to be measured in an order in which change of each control parameter becomes change corresponding to the change curves, and new measurement is performed such that only one of the control parameters changes from previous measurement.
    Type: Grant
    Filed: March 8, 2017
    Date of Patent: September 10, 2019
    Assignee: FUJITSU LIMITED
    Inventor: Yuhei Umeda