US 12,596,721 B1 · Domain-integrated contextual response engine in an artificial intelligence system

Abstract

Methods, systems, and computer storage media for providing domain-integrated contextual response management using a domain-integrated contextual response engine in an artificial intelligence (AI) system are described. Domain-integrated contextual response management is a systematic approach that combines specific industry knowledge with contextual understanding to generate accurate, relevant, and specific industry-tailored responses to user queries. Domain-integrated contextual response management further includes fine-tuning models for Retrieval-Augmented Generation (RAG) tasks using customer-specific data based on a two-fold approach involving skill distillation and knowledge distillation (i.e., skill distillation from a more powerful model like Large Language Model “LLM” and knowledge distillation from domain-specific data). Domain-integrated contextual response management also includes creating a synthetic dataset that enables smaller models (e.g., domain-integrated contextual response models) to effectively manage RAG tasks while incorporating domain-specific knowledge. Domain-integrated contextual response management further ensures that the domain-integrated contextual response models can retrieve relevant information, support citations, and decline out-of-domain (OOD) questions.

Claims20

Dependent claims

A computerized system comprising:

one or more computer processors;
and
computer memory storing computer-useable instructions that, when used by the one or more computer processors, cause the one or more computer processors to perform operations, the operations comprising:
accessing synthetic data comprising questions, the synthetic data is generated based on a synthetic data generator engine, wherein the questions include in-domain questions and out-of-domain questions;
generating a first training dataset, wherein generating the first training dataset is based on answering the in-domain questions and the out-of-domain questions in the synthetic data, wherein answering the in-domain questions and the out-of-domain questions is based on citation-and-filtration operations that provide citations for the in-domain questions and decline answering out-of-domain questions;
generating a second training dataset, wherein generating the second training dataset is based on one or more paraphrased versions of the questions in the synthetic data;
and
training a domain-integrated contextual response model based on the first training dataset and the second training dataset.

5 more elements

One or more computer-storage media having computer-executable instructions embodied thereon that, when executed by a computing system having a processor and memory, cause the processor to perform operations, the operations comprising:

communicating, a query from a client associated with an artificial intelligence (AI) agent;
based on communicating the query, receiving a response associated with the query, wherein the response is generated based on a domain-integrated contextual response model that is trained using a synthetic data generator engine, a first training dataset, a second training dataset and citation-and-filtration operations, wherein the first training dataset is generated based on questions including in-domain questions and out-of-domain questions, the second training dataset is generated based on paraphrased versions of the questions, and the citation-and-filtration operations provide citations for the in-domain questions and decline answer out-of-domain questions;
and
causing display of the response to the query.

1 more element

Dependencies are inferred from the claim text.

Domain-integrated contextual response engine in an artificial intelligence system

Life of this patent

Domain-integrated contextual response engine in an artificial intelligence system

Life of this patent

Abstract

Claims20