-
公开(公告)号:US11295234B2
公开(公告)日:2022-04-05
申请号:US15854167
申请日:2017-12-26
Applicant: Rovi Guides, Inc.
Inventor: Sankar Ardhanari , Sai Rahul Reddy Pulikunta , Sashikumar Venkataraman , Abubakkar Siddiq , Ganesh Ramamoorthy
IPC: G06N20/00 , G06K9/62 , G06F40/205 , H04N21/482 , H04N21/84
Abstract: Methods and systems are disclosed herein for accurately training a machine learning model with a reduced training data set. A large number of data records may be parsed. Each record may be reduced to a set of symbols representing the composition of each record. A user may assign a classification to each symbol within each record. Records with identical arrangements and classifications of symbols may be grouped together, and a representative sample of data records from each group may be fed into the model as the reduced training data set.
-
公开(公告)号:US10162853B2
公开(公告)日:2018-12-25
申请号:US15056828
申请日:2016-02-29
Applicant: Rovi Guides, Inc.
IPC: G06F17/30 , G10L15/18 , G10L15/22 , G10L13/04 , G10L13/033 , H04N21/482 , H04N21/258
Abstract: Methods and systems are described for responding to natural language queries. Specifically a natural language query is received and a query template that corresponds to the natural language query is determined. One or more search results corresponding to the natural language query are retrieved. One or more attributes associated with the user equipment device are identified and based on those attributes a first response template associated with an audio-only response to the natural language query, a second response template associated with a visual-only response to the natural language query, or a third response template associated with an audio-visual response to the natural language query is selected. A response to the natural language query is generated based on the selected response template and the retrieved one or more search results.
-
公开(公告)号:US20170161319A1
公开(公告)日:2017-06-08
申请号:US15056771
申请日:2016-02-29
Applicant: Rovi Guides, Inc.
IPC: G06F17/30 , H04N21/482
CPC classification number: G06F16/243 , G06F16/24575 , G06F16/24578 , G06F16/9535 , G10L15/22 , H04N21/25833 , H04N21/25841 , H04N21/4828
Abstract: Methods and systems are provided for generating a response to a natural language query. Specifically, the natural language query is received and a query template that corresponds to the natural language query is determined. One or more search results corresponding to the natural language query are retrieved. One or more attributes associated with a user are selected based on selection criteria. Based on those attributes, a response template previously assigned to the query template is identified. A response to the natural language query is generated based on the identified response template and the retrieved one or more search results.
-
-