Machine learning based disambiguation in a knowledge aware conversation system

A system for machine learning based disambiguation in a knowledge aware conversation system, comprising:

• one or more processors;
• and
• a memory comprising instructions that, when executed by the one or more processors, cause the system to:
• receive a query via a user interface;
• provide, as inputs to a given language processing machine learning model:
• the query;
• conversation history data associated with the query;
• and
• a prompt instructing the given language processing machine learning model to generate an enriched query by rewording the query based on conversation history data;
• generate, by the given language processing machine learning model based on the inputs, the enriched query, wherein the generating of the enriched query comprises extracting relevant contextual information from the conversation history data using the given language processing machine learning model, wherein the given language processing machine learning model uses the extracted relevant contextual information to reword the query;
• retrieve relevant information from a data store based on using an embedding of the enriched query to perform a semantic search;
• provide the enriched query and the relevant information to a language processing machine learning model along with a given prompt that instructs the language processing machine learning model to:
• generate an answer to the enriched query based on the relevant information if the language processing machine learning model detects no ambiguity in the enriched query;
• and
• generate a disambiguation question if the language processing machine learning model detects an ambiguity in the enriched query;
• receive an output from the language processing machine learning model in response to the given prompt, wherein the output comprises the answer or the disambiguation question;
• and
• provide a response to the query via the user interface based on the output.

A non-transitory computer readable medium comprising instructions that, when executed by one or more processors of a computing system, cause the computing system to:

• receive a query via a user interface;
• provide, as inputs to a given language processing machine learning model:
• the query;
• conversation history data associated with the query;
• and
• a prompt instructing the given language processing machine learning model to generate an enriched query by rewording the query based on the conversation history data;
• generate, by the given language processing machine learning model based on the inputs, the enriched query, wherein the generating of the enriched query comprises extracting relevant contextual information from the conversation history data using the given language processing machine learning model, wherein the given language processing machine learning model uses the extracted relevant contextual information to reword the query;
• retrieve relevant information from a data store based on using an embedding of the enriched query to perform a semantic search;
• provide the enriched query and the relevant information to a language processing machine learning model along with a given prompt that instructs the language processing machine learning model to:
• generate an answer to the enriched query based on the relevant information if the language processing machine learning model detects no ambiguity in the enriched query;
• and
• generate a disambiguation question if the language processing machine learning model detects an ambiguity in the enriched query;
• receive an output from the language processing machine learning model in response to the given prompt, wherein the output comprises the answer or the disambiguation question;
• and
• provide a response to the query via the user interface based on the output.