Providing Additional Information for Identified Named-Entities for Assistant Systems

Abstract

In one embodiment, a method includes receiving, from a first client system associated with a first user, a message sent from the first user to a second user comprising one or more n-grams; automatically tagging, responsive to identifying the one or more named-entities, one or more of the n-grams of the message with references to the one or more identified named-entities; sending, to a second client system associated with the second user, instructions for presenting the message to the second user, where the message comprises the one or more tagged n-grams corresponding to the one or more identified named-entities, where each tagged n-gram is selectable to execute a task associated with the corresponding identified named-entity; receiving, from the second client system, an indication that the second user selected one or more tagged n-grams to execute the task associated with the corresponding identified named-entity; and sending, to the second client system, instructions for presenting results of the executed task.

What is claimed is: 
     
         1 . A method comprising, by one or more computing systems:
 receiving, from a first client system associated with a first user, a message sent from the first user to a second user, wherein the message comprises one or more n-grams;   automatically tagging, responsive to identifying the one or more named-entities, one or more of the n-grams of the message with references to the one or more identified named-entities;   sending, to a second client system associated with the second user, instructions for presenting the message to the second user, wherein the message comprises the one or more tagged n-grams corresponding to the one or more identified named-entities, wherein each tagged n-gram is selectable to execute a task associated with the corresponding identified named-entity;   receiving, from the second client system, an indication that the second user selected one or more tagged n-grams to execute the task associated with the corresponding identified named-entity; and   sending, to the second client system, instructions for presenting results of the executed task.   
     
     
         2 . The method of  claim 1 , further comprising:
 analyzing, by the one or more computing systems, the received message to identify one or more named-entities corresponding to one or more of the n-grams, wherein each of the one or more named-entities is associated with a confidence score higher than a pre-determined score value representing a likelihood that the named-entity is referenced by the n-gram in the message.   
     
     
         3 . The method of  claim 2 , wherein analyzing the received message comprises:
 identifying one or more slots in the message by performing a semantic parsing on the message or by processing the message with a natural-language processing (NLP) algorithm;   identifying, for each of the one or more identified slots, one or more named-entities referenced by the slot by performing entity resolution algorithms for the slot in the received message, wherein each of the one or more named-entities is associated with a confidence score representing a likelihood that the named-entity is referenced by the corresponding slot in the message; and   verifying, for each of the one or more identified slots, one or more named-entities are referenced by the slot in the message.   
     
     
         4 . The method of  claim 3 , wherein performing entity resolution algorithms for the slot comprises:
 identifying one or more candidate named-entities that may be referenced by the slot;   computing, for each of the one or more candidate named-entities, a confidence score based on a context of the message; and   selecting, if a highest computed confidence score is higher than the other computed confidence scores by more than a pre-determined score value, a candidate named-entity associated with the highest confidence score as a named-entity referenced by the slot.   
     
     
         5 . The method of  claim 3 , wherein performing entity resolution algorithms for the slot comprises:
 identifying a plurality of candidate named-entities that may be referenced by the slot;   computing, for each of the plurality of candidate named-entities, a confidence score based on a context of the message; and   selecting, if a highest computed confidence score is not higher than one or more the other computed confidence scores by at least a pre-determined score value, a subset of the plurality of candidate named-entities as named-entities potentially referenced by the slot, wherein a difference between the highest computed confidence score and a computed confidence score corresponding to each named-entity in the subset is less than the pre-determined score value.   
     
     
         6 . The method of  claim 4 , further comprising:
 receiving, at the one or more computing systems from the second client system associated with the second user, an indication that the second user selected one of the one or more tagged n-grams, wherein the selected tagged n-gram is associated with the named-entities in the subset; and   sending, from the one or more computing systems to the second client system associated with the second user, instructions for presenting a list of the named-entities in the subset to the second user, wherein each named-entity in the list is selectable to retrieve additional information associated with the named-entity.   
     
     
         7 . The method of  claim 6 , further comprising:
 receiving, at the one or more computing systems from the second client system associated with the second user, an indication that the second user selected a particular named-entity from the named-entities in the list;   retrieving, by the one or more computing systems, additional information associated with the particular named-entity from a knowledge graph; and   sending, from the one or more computing systems to the second client system associated with the second user, instructions for presenting the retrieved additional information associated with the particular named-entity to the second user.   
     
     
         8 . The method of  claim 7 , further comprising:
 updating user behavior history records associated with the second user based on the received indication from the second user, wherein the user behavior history records comprise information on the list of named entities presented to the second user and the selected particular named entity; and   storing the updated user behavior history records in association with a user profile of the second user, wherein the updated user behavior history records will be used for future entity resolutions.   
     
     
         9 . The method of  claim 2 , wherein verifying one or more named-entities are referenced by the slot in the message comprises determining whether a highest confidence score associated with one of the one or more identified named-entities is higher than a threshold score. 
     
     
         10 . The method of  claim 1 , further comprising:
 receiving, at the one or more computing systems from the second client system associated with the second user, an indication that the second user selected one of the one or more tagged n-grams, wherein the selected tagged n-gram is associated with a particular named-entity;   retrieving by the one or more computing systems, additional information associated with the particular named-entity from a knowledge graph; and   sending, from the one or more computing systems to the second client system associated with the second user, instructions for presenting the retrieved additional information associated with the particular named-entity to the second user.   
     
     
         11 . The method of  claim 10 , further comprising:
 updating user behavior history records associated with the second user based on the received indication from the second user, wherein the user behavior history records comprise information on the particular named-entity associated with the selected n-gram; and   storing the updated user behavior history records in association with a user profile of the second user.   
     
     
         12 . The method of  claim 1 , wherein each tagged n-gram is underlined or highlighted to indicate the tagged n-gram is selectable. 
     
     
         13 . The method of  claim 1 , wherein a button is displayed close to each tagged n-gram to indicate the tagged n-gram is selectable. 
     
     
         14 . The method of  claim 1 , wherein the task is subject to the second user's privacy settings. 
     
     
         15 . One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
 receive, from a first client system associated with a first user, a message sent from the first user to a second user, wherein the message comprises one or more n-grams;   automatically tag, responsive to identifying the one or more named-entities, one or more of the n-grams of the message with references to the one or more identified named-entities;   send, to a second client system associated with the second user, instructions for presenting the message to the second user, wherein the message comprises the one or more tagged n-grams corresponding to the one or more identified named-entities, wherein each tagged n-gram is selectable to execute a task associated with the corresponding identified named-entity;   receive, from the second client system, an indication that the second user selected one or more tagged n-grams to execute the task associated with the corresponding identified named-entity; and   send, to the second client system, instructions for presenting results of the executed task.   
     
     
         16 . The media of  claim 15 , wherein to analyze the received message comprises steps to:
 identify one or more slots in the message by performing a semantic parsing on the message or by processing the message with a natural-language processing (NLP) algorithm;   identify, for each of the one or more identified slots, one or more named-entities referenced by the slot by performing entity resolution algorithms for the slot in the received message, wherein each of the one or more named-entities is associated with a confidence score representing a likelihood that the named-entity is referenced by the corresponding slot in the message; and   verify, for each of the one or more identified slots, one or more named-entities are referenced by the slot in the message.   
     
     
         17 . The media of  claim 16 , wherein to perform entity resolution algorithms for the slot comprises steps to:
 identify one or more candidate named-entities that may be referenced by the slot;   compute, for each of the one or more candidate named-entities, a confidence score based on a context of the message; and   select, if a highest computed confidence score is higher than the other computed confidence scores by more than a pre-determined score value, a candidate named-entity associated with the highest confidence score as a named-entity referenced by the slot.   
     
     
         18 . The media of  claim 16 , wherein to perform entity resolution algorithms for the slot comprises step to:
 identify a plurality of candidate named-entities that may be referenced by the slot;   compute, for each of the plurality of candidate named-entities, a confidence score based on a context of the message; and   select, if a highest computed confidence score is not higher than one or more the other computed confidence scores by at least a pre-determined score value, a subset of the plurality of candidate named-entities as named-entities potentially referenced by the slot, wherein a difference between the highest computed confidence score and a computed confidence score corresponding to each named-entity in the subset is less than the pre-determined score value.   
     
     
         19 . The media of  claim 18 , wherein the software is further operable when executed to:
 receive, from the second client system associated with the second user, an indication that the second user selected one of the one or more tagged n-grams, wherein the selected tagged n-gram is associated with the named-entities in the subset; and   send, to the second client system associated with the second user, instructions for presenting a list of the named-entities in the subset to the second user, wherein each named-entity in the list is selectable to retrieve additional information associated with the named-entity.   
     
     
         20 . A system comprising: one or more processors; and a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
 receive, from a first client system associated with a first user, a message sent from the first user to a second user, wherein the message comprises one or more n-grams;   automatically tag, responsive to identifying the one or more named-entities, one or more of the n-grams of the message with references to the one or more identified named-entities;   send, to a second client system associated with the second user, instructions for presenting the message to the second user, wherein the message comprises the one or more tagged n-grams corresponding to the one or more identified named-entities, wherein each tagged n-gram is selectable to execute a task associated with the corresponding identified named-entity;   receive, from the second client system, an indication that the second user selected one or more tagged n-grams to execute the task associated with the corresponding identified named-entity; and   send, to the second client system, instructions for presenting results of the executed task.