System and method for combining unsupervised machine learning and information extraction models for topic modeling

A system comprising:

• a memory having computer-readable instructions stored thereon;
• and
• a processor that executes the computer-readable instructions to:
• receive a set of documents from which to generate at least one topic;
• create at least one information extraction model for the set of documents wherein creating the at least one information extraction model includes:
• identifying key information associated with a domain to which the set of documents belong to, wherein identifying the key information associated with the domain includes:
• identifying one or more grammatical patterns from the set of documents;
• identifying an entity type from the set of documents;
• identifying one or more of a keyword, phrase, or list of information from the set of documents;
• and
• creating the one or more rule sets based on the one or more grammatical patterns, the entity type, and the one or more of the keyword, phrase, or list of information by:
• defining one or more rules for extracting first text segments associated with the one or more grammatical patterns, the entity type, and the one or more of the keyword, phrase, or list of information;
• and
• adding context around the first text segments to obtain the one or more rule sets;
• and
• encoding the key information into the at least one information extraction model;
• execute the at least one information extraction model to extract a plurality of text segments from the set of documents, wherein extracting the plurality of text segments comprises:
• applying one or more rule sets defined for at least one entity, word list, or grammatical pattern;
• and
• extracting a plurality of text portions from the set of documents based on the one or more rule sets as the plurality of text segments;
• input at least a subset of the plurality of text segments into an unsupervised machine learning model;
• and
• execute the unsupervised machine learning model to output the at least one topic for the set of the documents,
• wherein:
• the at least one topic is used to generate a compressed prompt representing the set of documents,
• generating the compressed prompt includes concatenating a plurality of topic terms output by the unsupervised machine learning model,
• the plurality of topic terms being concatenated in an order from a given topic term having a highest weight value to another given topic term having a lowest weight value,
• a given weight value assigned to a topic term of the plurality of topic terms represents a relevance of the topic term to the at least one topic;
• the compressed prompt is input into a language model that is executed to generate a summary of the set of documents.

A method comprising:

• receiving, by a processor executing computer-readable instructions stored on a memory, a set of documents from which to generate at least one topic;
• creating, by the processor, at least one information extraction model for the set of documents, wherein creating the at least one information extraction model includes:
• identifying key information associated with a domain to which the set of documents belong to, wherein identifying the key information associated with the domain includes:
• identifying one or more grammatical patterns from the set of documents;
• identifying an entity type from the set of documents;
• identifying one or more of a keyword, phrase, or list of information from the set of documents;
• and
• creating the one or more rule sets based on the one or more grammatical patterns, the entity type, and the one or more of the keyword, phrase, or list of information by:
• defining one or more rules for extracting first text segments associated with the one or more grammatical patterns, the entity type, and the one or more of the keyword, phrase, or list of information;
• and
• adding context around the first text segments to obtain the one or more rule sets;
• and
• encoding the key information into the at least one information extraction model;
• executing, by the processor, the at least one information extraction model to extract a plurality of text segments from the set of documents, wherein extracting the plurality of text segments comprises:
• applying, by the processor, one or more rule sets defined for at least one entity, word list, or grammatical pattern;
• and
• extracting, by the processor, a plurality of text portions from the set of documents based on the one or more rule sets as the plurality of text segments;
• inputting, by the processor, at least a subset of the plurality of text segments into an unsupervised machine learning model;
• and
• executing, by the processor, the unsupervised machine learning model to output the at least one topic for the set of the documents,
• wherein:
• the at least one topic is used to generate a compressed prompt representing the set of documents,
• generating the compressed prompt includes concatenating a plurality of topic terms output by the unsupervised machine learning model,
• the plurality of topic terms being concatenated in an order from a given topic term having a highest weight value to another given topic term having a lowest weight value,
• a given weight value assigned to a topic term of the plurality of topic terms represents a relevance of the topic term to the at least one topic;
• and
• the compressed prompt is input into a language model that is executed to generate a summary of the set of documents.