公开(公告)号：US10788900B1

公开(公告)日：2020-09-29

申请号：US16023912

申请日：2018-06-29

申请人： Snap Inc.

发明人： William Brendel ,  Francesco Barbieri ,  Xin Chen ,  Wei Chu ,  Venkata Satya Pradeep Karuturi ,  Luis Carlos Dos Santos Marujo ,  Leonardo Ribas Machado das Neves

IPC分类号： G06F17/27 ,  G06F3/023 ,  G06N3/08 ,  G06K9/62 ,  G06F3/0481 ,  H04L12/58 ,  G06F40/166 ,  G06F40/274

摘要： Symbol prediction can be implemented using a multi-task system trained for different tasks. The tasks may include a single symbol prediction, symbol category prediction, and symbol subcategory prediction. Categories of symbols can be generated by clustering sets of training data using a clustering scheme.

公开(公告)号：US11620001B2

公开(公告)日：2023-04-04

申请号：US16948018

申请日：2020-08-27

申请人： Snap Inc.

发明人： William Brendel ,  Francesco Barbieri ,  Xin Chen ,  Wei Chu ,  Venkata Satya Pradeep Karuturi ,  Luis Carlos Dos Santos Marujo ,  Leonardo Ribas Machado das Neves

IPC分类号： G06F40/166 ,  G06F3/023 ,  G06N3/084 ,  G06K9/62 ,  G06F3/04817 ,  H04L51/04 ,  G06F40/274

摘要： Symbol prediction can be implemented using a multi-task system trained for different tasks. The tasks may include a single symbol prediction, symbol category prediction, and symbol subcategory prediction. Categories of symbols can be generated by clustering sets of training data using a clustering scheme.