-
11.发明公开公开(公告)号：US20240256923A1
公开(公告)日：2024-08-01
申请号：US18415906
申请日：2024-01-18
Applicant: Fujitsu Limited , RIKEN
Inventor： Mutsuyo WADA , Yuichiro WADA , Kimihiro YAMAZAKI , Takashi KATOH , Atsushi TOKUHISA
IPC: G06N5/04
CPC classification number: G06N5/04
Abstract: A non-transitory computer-readable recording medium stores therein a path selecting program that causes a computer to execute a process including extracting a plurality of representative points from a presence probability distribution of states of a target, identifying a first plurality of state transition paths between the plurality of representative points, and selecting a second plurality of state transition paths from the first plurality of state transition paths, based on a probability density of each path included in the first plurality of state transition paths.
-
公开(公告)号：US20240176583A1
公开(公告)日：2024-05-30
申请号：US18370436
申请日：2023-09-20
Applicant: Fujitsu Limited
Inventor： Takashi KATOH
Abstract: For managing an order of pieces of data, a computer stores, in order, a plurality of pieces of value data that represent a number of digits with which a constant multiple of a number of pieces of data included in a data group are representable; stores flag defining an upper-lower relationship between first and second ranges in each of which at least the number of pieces of data is representable; each time designation of first value data to be moved to a head in the order is accepted, sets a value that belongs to an upper range among the first and the second ranges to the first value data such that the first value data is before other pieces of value data in the order; and when the lower range no longer includes any pieces of value data, updates the flag such that the upper-lower relationship is switched.
-
公开(公告)号：US20240176522A1
公开(公告)日：2024-05-30
申请号：US18507155
申请日：2023-11-13
Applicant: Fujitsu Limited
Inventor： Takashi KATOH
IPC: G06F3/06
CPC classification number: G06F3/064 , G06F3/0688 , G06F3/0604
Abstract: An information processing program causes a computer to store a plurality of block strings that represents arrays in which blocks associated with a predetermined number of components of an array are coupled. For each block string, the computer manages the block string such that, when there is a front block corresponding only to components for which a value has not been set in a first half portion of the block string, at least a plurality of areas that is a part of a last block is unused areas where values of components are not stored, by distributing values of a rear block in a latter half portion to the front block. A pointer indicating a position at which the block string is divided and initial values of components are also managed by using a second area of the plurality of areas in the last block of the each block string.
-
公开(公告)号：US20230376846A1
公开(公告)日：2023-11-23
申请号：US18199443
申请日：2023-05-19
Applicant: Fujitsu Limited , INDIAN INSTITUTE OF SCIENCE
Inventor： Sho TAKEMORI , Takashi KATOH , Yuhei UMEDA , Harsh RANGWANI , Shrinivas RAMASUBRAMANIAN , Venkatesh Babu RADHAKRISHNAN
Abstract: An information processing apparatus includes one or more memories; and one or more processors coupled to the one or more memories, the one or more processors being configured to decide a gain matrix based on an input metric, perform selection of first training data from a plurality of unlabeled training data, to be used for training a machine learning model, based on the gain matrix, and perform training of the machine learning model based on the first training data, a predicted label that is predicted from the first training data, and a loss function including the gain matrix.
-
15.发明公开公开(公告)号：US20230289659A1
公开(公告)日：2023-09-14
申请号：US18174625
申请日：2023-02-26
Applicant: Fujitsu Limited
Inventor： Takashi KATOH , Kento UEMURA , Suguru YASUTOMI
IPC: G06N20/00
CPC classification number: G06N20/00
Abstract: An information processing method comprising: for a classification model for classifying input data into one or another of plural classes that was trained using a first data set, identifying, in a second data set that is different from the first data set one or more items of data having a specific datum of which a degree of contribution to a change in a classification criterion is greater than a predetermined threshold, the classification criterion being a classification criterion of the classification model during re-training based on the second data set; and, from among the one or more items of data, detecting an item of data, for which a loss reduces for the classification model by change to the classification criterion by re-training based on the second data set, as an item of data of an unknown class not contained in the plural classes.
-
Patent Agency Ranking
公开(公告)号：US20190286937A1
公开(公告)日：2019-09-19
申请号：US16287494
申请日：2019-02-27
Applicant: FUJITSU LIMITED
Inventor： Takashi KATOH , Kento UEMURA , Suguru YASUTOMI
Abstract: A learning device executes learning of a discriminator that discriminates object data to a known class included in training data or an unknown class not included in the training data, using the training data. The learning device then generates a feature value of the unknown class, from a feature value of a plurality of layers of the discriminator, by at least a part of the training data in the layers. The learning device then executes the learning of the discriminator so that a feature value of the known class and the generated feature value of the unknown class are separated.
-
公开(公告)号：US20230289624A1
公开(公告)日：2023-09-14
申请号：US18088825
申请日：2022-12-27
Applicant: Fujitsu Limited
Inventor： Takashi KATOH , Kento UEMURA , Suguru YASUTOMI
IPC: G06N5/022
CPC classification number: G06N5/022
Abstract: A non-transitory computer-readable storage medium storing an information processing program that causes at least one computer to execute a process, the process includes acquiring an update amount of a classification criterion of a classification model in retraining, the classification model being trained by using a first dataset, the classification model classifying input data into one of a plurality of classes, the retraining being performed by using a second dataset; and detecting data with a largest change amount among the second dataset when changing each piece of data included in the second dataset so as to decrease the update amount.
-
18.发明公开公开(公告)号：US20230281845A1
公开(公告)日：2023-09-07
申请号：US18060338
申请日：2022-11-30
Applicant: Fujitsu Limited
Inventor： Suguru YASUTOMI , Akira SAKAI , Takashi KATOH , Kento UEMURA
CPC classification number: G06T7/507 , G06T5/50 , G06T7/60 , G06V10/758 , G06T2207/10132 , G06T2207/20221
Abstract: An information processing device acquires output image data that is acquired by inputting image data indicating a pseudo-shadow area to an auto-encoder that is generated by machine learning using label image data contained in training data, the label image data indicating a shadow area in ultrasound image data of a captured target. The information processing device generates augmented data corresponding to the training data by combining the acquired output image data with the ultrasound image data.
-
公开(公告)号：US20220147764A1
公开(公告)日：2022-05-12
申请号：US17473509
申请日：2021-09-13
Applicant: FUJITSU LIMITED
Inventor： Takashi KATOH , Kento UEMURA , Suguru YASUTOMI , Tomohiro HAYASE
Abstract: A non-transitory computer-readable storage medium storing a data generation program that causes at least one computer to execute a process, the process includes, acquiring a data generation model that is trained by using a first dataset corresponding to a first domain and a second dataset corresponding to a second domain, and that includes an identification loss by an identification model in a parameter; inputting first data corresponding to the first domain to the identification model to acquire a first identification loss, and inputting second data corresponding to the second domain to the identification model to acquire a second identification loss; generating data in which the second identification loss approximates the first identification loss, by using the data generation model; and outputting the data that is generated.
-
公开(公告)号：US20220101124A1
公开(公告)日：2022-03-31
申请号：US17368890
申请日：2021-07-07
Applicant: FUJITSU LIMITED
Inventor： Suguru YASUTOMI , Tomohiro HAYASE , Takashi KATOH
Abstract: A non-transitory computer-readable storage medium storing an information processing program that causes at least one computer to execute a process, the process includes acquiring a first machine learning model trained by using a training data set including first data and a second machine learning model not trained with the specific data; and retraining the first machine learning model so that an output of the first machine learning model and an output of the second machine learning model when second data corresponding to the first data is input get close to each other.
-
-
-
-
-
-
-
-
-
Search Results
20
records
(took 0.015 seconds)
