Modifying dynamic conversational responses based on detecting references to socially close people
AssigneeCapital One Services, LLC
InventorsAlexandra Coman, Chihyen Yang, Rui Zhang, Jihoon Jay Song
Life of this patent
Estimating the term…
Abstract
Methods and systems are described for generating dynamic conversational responses sensitive to different emotional contexts using machine learning models. The dynamic conversational responses may be generated in real time and reflect the likely emotional context by detecting socially close entities and events in user input.
Claims20
A system for generating dynamic conversational responses sensitive to different emotional contexts using machine learning models, the system comprising:
- one or more processors;
- and
- a non-transitory computer-readable storage medium storing instructions, which, when executed by the one or more processors, cause the one or more processors to:
- receive a first user input from a user during a conversational interaction with a user interface of a device associated with the user;
- compare a plurality of terms within the first user input with a plurality of stored terms, wherein each term of the plurality of stored terms is associated with a corresponding social closeness value, and wherein a particular social closeness value indicates a social distance between the user and an entity represented by a corresponding stored term;
- based on comparing the plurality of terms within the first user input with the plurality of stored terms, determine a term of the plurality of terms that matches a stored term of the plurality of stored terms;
- retrieve a social closeness value associated with the term of the plurality of terms;
- in response to determining that the social closeness value is above a threshold, determine an emotional context of the first user input;
- generate a feature input based on the first user input, the term of the plurality of terms, the emotional context, and the social closeness value;
- input the feature input into a first machine learning model, wherein the first machine learning model is trained to select or generate a dynamic conversational response from a plurality of dynamic conversational responses based on the feature input;
- and
- generate for display, at the user interface, the dynamic conversational response selected by the first machine learning model during the conversational interaction.
A method for generating dynamic conversational responses sensitive to different emotional contexts using machine learning models, the method comprising:
- receiving a first user input from a user during a conversational interaction with a user interface;
- identifying, in the first user input, a term that matches a stored term with an associated social closeness value, wherein the associated social closeness value indicates a social distance between the user and an entity represented by the stored term;
- determining, using the first user input, an emotional context associated with the user;
- generating a feature input based on the first user input, the term, the emotional context, and the associated social closeness value;
- inputting the feature input into a first machine learning model to obtain a dynamic conversational response from a plurality of dynamic conversational responses based on the feature input;
- and
- generating for display, at the user interface, the dynamic conversational response based on a response selected by the first machine learning model during the conversational interaction.
A non-transitory, computer readable medium for generating dynamic conversational responses using machine learning models, comprising instructions that, when executed by one or more processors, cause operations comprising:
- receiving a first user input during a conversational interaction by a user with a user interface;
- identifying, in the first user input, a term that matches a stored term with an associated social closeness value, wherein the associated social closeness value indicates a social distance between the user and an entity represented by the stored term;
- determining, using the first user input, an emotional context associated with the user;
- generating a feature input based on the first user input, the term, the emotional context, and the associated social closeness value;
- inputting the feature input into a first machine learning model to obtain a dynamic conversational response from a plurality of dynamic conversational responses based on the feature input;
- and
- generating for display, at the user interface, the dynamic conversational response based on a response selected by the first machine learning model during the conversational interaction.
Dependencies are inferred from the claim text.