公开(公告)号：US20210409789A1

公开(公告)日：2021-12-30

申请号：US16612498

申请日：2018-09-28

Applicant: KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY

Inventor： Dongsu HAN ,  Hyunho YEO ,  Youngmok JUNG ,  Jaehong KIM ,  Jinwoo SHIN

IPC: H04N21/2383 ,  H04L29/08 ,  G06N3/04 ,  G06N3/08

Abstract: A method and apparatus for transmitting adaptive video in real time using a content-aware neural network are disclosed. At least one embodiment provides a method performed by a server for transmitting an adaptive video in real time by using content-aware deep neural networks (DNNs), including downloading a video, encoding a downloaded video for each of at least one resolution, dividing an encoded video into video chunks of a predetermined size, training the content-aware DNNs by using encoded video, generating a configuration or manifest file containing information on trained content-aware DNNs and information on the encoded video, and transmitting the configuration file upon a request of a client.

公开(公告)号：US20210374536A1

公开(公告)日：2021-12-02

申请号：US17334204

申请日：2021-05-28

Applicant: SAMSUNG ELECTRONICS CO., LTD. ,  KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY

Inventor： Youyoung SONG ,  Seungwoo SEO ,  Jinwoo SHIN ,  Eunho YANG ,  Sungju HWANG

IPC: G06N3/08 ,  G06K9/62 ,  G06F16/901 ,  G06F16/903

Abstract: A method of training a retrosynthesis prediction model includes determining first attention information from first character string information of a product, based on first graph information of the product, encoding the first character string information, based on the determined first attention information, and determining second attention information from the first graph information and second graph information of a reactant. The method further includes decoding second character string information of the reactant, based on the determined second attention information, and training the retrosynthesis prediction model, based on the decoded second character string information.

公开(公告)号：US20220392051A1

公开(公告)日：2022-12-08

申请号：US17574959

申请日：2022-01-13

Applicant: Samsung Electronics Co., Ltd. ,  Korea Advanced Institute of Science and Technology

Inventor： Youngbum HUR ,  Jinwoo SHIN ,  Jihoon TACK

IPC: G06T7/00 ,  G06V10/82 ,  G06T11/00 ,  G06V10/77 ,  G06V10/40 ,  G06T3/60 ,  G06V10/774 ,  G06V10/764

Abstract: A processor-implemented method with image analysis includes: receiving a test image; generating a plurality of augmented images by augmenting the test image; determining classification prediction values for the augmented images using a classifier; determining a detection score based on the classification prediction values; and determining whether the test image corresponds to anomaly data based on the detection score and a threshold.

公开(公告)号：US20190122081A1

公开(公告)日：2019-04-25

申请号：US15798237

申请日：2017-10-30

Applicant: Korea Advanced Institute of Science and Technology

Inventor： Jinwoo SHIN ,  Kimin LEE

IPC: G06K9/62 ,  G06N3/08

Abstract: Disclosed herein are a confident deep learning ensemble method and apparatus based on specialization. In one aspect, a confident deep learning ensemble method based on specialization proposed by the present invention includes the steps of generating a target function of maximizing entropy by minimizing Kullback-Leibler divergence with a uniform distribution with respect to the not-classified data of models for image processing and generating general features by sharing features between the models and performing learning for image processing using the general features.