公开(公告)号：US20230385551A1

公开(公告)日：2023-11-30

申请号：US18201075

申请日：2023-05-23

Applicant: Snap Inc.

Inventor： Di Lu ,  Leonardo Ribas Machado das Neves ,  Vitor Rocha de Carvalho ,  Ning Zhang

IPC: G06F40/295 ,  G06N20/00 ,  G06N3/08 ,  G06F40/30

CPC classification number: G06F40/295 ,  G06N20/00 ,  G06N3/08 ,  G06F40/30

Abstract: A caption of a multimodal message (e.g., social media post) can be identified as a named entity using an entity recognition system. The entity recognition system can use a visual attention based mechanism to generate a visual context representation from an image and caption. The system can use the visual context representation to identify one or more terms of the caption as a named entity.

公开(公告)号：US11750547B2

公开(公告)日：2023-09-05

申请号：US17459161

申请日：2021-08-27

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Leonardo Ribas Machado das Neves ,  Seungwhan Moon

IPC: H04L51/10 ,  G06N3/08 ,  G06V30/19 ,  G06V10/82 ,  G06V10/40 ,  H04L67/10

CPC classification number: H04L51/10 ,  G06N3/08 ,  G06V10/40 ,  G06V10/82 ,  G06V30/19147 ,  G06V30/19173 ,  H04L67/10

Abstract: A caption of a multimodal message (e.g., social media post) can be identified as a named entity using an entity recognition system. The entity recognition system can use an attention-based mechanism that emphasis or de-emphasizes each data type (e.g., image, word, character) in the multimodal message based on each datatypes relevance. The output of the attention mechanism can be used to update a recurrent network to identify one or more words in the caption as being a named entity.

公开(公告)号：US20240354508A1

公开(公告)日：2024-10-24

申请号：US18761154

申请日：2024-07-01

Applicant: Snap Inc.

Inventor： Di Lu ,  Leonardo Ribas Machado das Neves ,  Vitor Rocha de Carvalho ,  Ning Zhang

IPC: G06F40/295 ,  G06F40/30 ,  G06N3/08 ,  G06N20/00

CPC classification number: G06F40/295 ,  G06F40/30 ,  G06N3/08 ,  G06N20/00

Abstract: A caption of a multimodal message (e.g., social media post) can be identified as a named entity using an entity recognition system. The entity recognition system can use a visual attention based mechanism to generate a visual context representation from an image and caption. The system can use the visual context representation to identify one or more terms of the caption as a named entity.

公开(公告)号：US20230013116A1

公开(公告)日：2023-01-19

申请号：US17945860

申请日：2022-09-15

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Leonardo Ribas Machado das Neves ,  Seungwhan Moon

IPC: G06K9/62 ,  G06N3/08

Abstract: A machine learning based system can identify an entity as the likely subject of a multimodal message (e.g., a social media post having a short text phrase overlaid on an image) by creating embeddings for an image of the multimodal message and one or more string embeddings from text of the multimodal message. The embeddings can be weighted to maximize information gain, then recombined and compared against a result embedding database to identify an entity as the subject of the multimodal message.

公开(公告)号：US12164603B2

公开(公告)日：2024-12-10

申请号：US17945860

申请日：2022-09-15

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Leonardo Ribas Machado das Neves ,  Seungwhan Moon

IPC: G06F18/25 ,  G06F18/21 ,  G06F18/22 ,  G06N3/08 ,  G06V30/10

Abstract: A machine learning based system can identify an entity as the likely subject of a multimodal message (e.g., a social media post having a short text phrase overlaid on an image) by creating embeddings for an image of the multimodal message and one or more string embeddings from text of the multimodal message. The embeddings can be weighted to maximize information gain, then recombined and compared against a result embedding database to identify an entity as the subject of the multimodal message.

公开(公告)号：US11704488B2

公开(公告)日：2023-07-18

申请号：US17544664

申请日：2021-12-07

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Luis Carlos Dos Santos Marujo ,  Leonardo Ribas Machado das Neves

IPC: G06F40/263 ,  G06F40/284 ,  G06F40/295 ,  G06N20/00

CPC classification number: G06F40/263 ,  G06F40/284 ,  G06F40/295 ,  G06N20/00

Abstract: Systems, devices, media, and methods are presented for generating a language detection model of a language analysis system. The systems and methods access a set of messages including text elements and convert the set of messages into a set of training messages. The set of training messages are configured for training a language detection model. The systems and methods train a classifier based on the set of training messages. The classifier has a set of features representing word frequency, character frequency, and a character ratio. The systems and methods generate a language detection model based on the classifier and the set of features.

公开(公告)号：US20220092261A1

公开(公告)日：2022-03-24

申请号：US17544664

申请日：2021-12-07

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Luis Carlos Dos Santos Marujo ,  Leonardo Ribas Machado das Neves

IPC: G06F40/263 ,  G06F40/284 ,  G06F40/295 ,  G06N20/00

Abstract: Systems, devices, media, and methods are presented for generating a language detection model of a language analysis system. The systems and methods access a set of messages including text elements and convert the set of messages into a set of training messages. The set of training messages are configured for training a language detection model. The systems and methods train a classifier based on the set of training messages. The classifier has a set of features representing word frequency, character frequency, and a character ratio. The systems and methods generate a language detection model based on the classifier and the set of features.

公开(公告)号：US11210467B1

公开(公告)日：2021-12-28

申请号：US15953357

申请日：2018-04-13

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Luis Carlos Dos Santos Marujo ,  Leonardo Ribas Machado das Neves

IPC: G06N20/00 ,  G06F40/263 ,  G06F40/295 ,  G06F40/284

Abstract: Systems, devices, media, and methods are presented for generating a language detection model of a language analysis system. The systems and methods access a set of messages including text elements and convert the set of messages into a set of training messages. The set of training messages are configured for training a language detection model. The systems and methods train a classifier based on the set of training messages. The classifier has a set of features representing word frequency, character frequency, and a character ratio. The systems and methods generate a language detection model based on the classifier and the set of features.

公开(公告)号：US11017173B1

公开(公告)日：2021-05-25

申请号：US16230341

申请日：2018-12-21

Applicant: Snap Inc.

Inventor： Di Lu ,  Leonardo Ribas Machado das Neves ,  Vitor Rocha de Carvalho ,  Ning Zhang

IPC: G06F40/295 ,  G06N20/00 ,  G06N3/08 ,  G06F40/30

Abstract: A caption of a multimodal message (e.g., social media post) can be identified as a named entity using an entity recognition system. The entity recognition system can use a visual attention based mechanism to generate a visual context representation from an image and caption. The system can use the visual context representation to identify one or more terms of the caption as a named entity.

公开(公告)号：US12155612B2

公开(公告)日：2024-11-26

申请号：US18358919

申请日：2023-07-25

Applicant: Snap Inc.

Inventor： Vitor Rocha de Carvalho ,  Leonardo Ribas Machado das Neves ,  Seungwhan Moon

IPC: H04L51/10 ,  G06N3/08 ,  G06V10/40 ,  G06V10/82 ,  G06V30/19 ,  H04L67/10

Abstract: A caption of a multimodal message (e.g., social media post) can be identified as a named entity using an entity recognition system. The entity recognition system can use an attention-based mechanism that emphasis or de-emphasizes each data type (e.g., image, word, character) in the multimodal message based on each datatypes relevance. The output of the attention mechanism can be used to update a recurrent network to identify one or more words in the caption as being a named entity.