User profiling using chain-of-thought knowledge graphs for querying a machine learning system

Patent metadata

Field	Value
Publication number	US-2026057254-A1
Application number	US-202418810039-A
Country	US
Kind code	A1
Filing date	Aug 20, 2024
Priority date	Aug 20, 2024
Publication date	Feb 26, 2026
Grant date	—

Abstract

Techniques are disclosed for a machine learning model, such as a large learning model (LLM), that incorporates a model of a chain of thought of a particular user when responding to a query from the user. In one example, a system generates a knowledge graph of a chain of thought of the user. The knowledge graph comprises nodes representing topics present within past queries by the user and edges representing a co-occurrence between the topics. The system determines, based on a topic present within a query from the user and the knowledge graph, a goal query comprising a goal topic. The system provides, to a machine learning model, the user to generate, by the machine learning model, a response. The machine learning model is constrained to include the goal topic of the goal query within the response. The system outputs, for display, the response to the query.

First claim

What is claimed is:

1 . A system comprising:
processing circuitry in communication with storage media, the processing circuitry configured to:
process records of past queries by a user to generate a knowledge graph corresponding to the user, the knowledge graph comprising nodes and edges, each of the nodes representing a topic of topics present within the past queries by the user, and each of the edges representing a co-occurrence between the topics present within the past queries by the user;
determine, from a query by the user, a topic present within the query;
determine, based at least in part on the topic present within the query and the knowledge graph corresponding to the user, a goal query, the goal query comprising a goal topic;
provide, to a machine learning model, the query by the user to generate, by the machine learning model, a response to the query by the user, wherein the goal topic constrains the machine learning model to include the goal topic of the goal query within the response; and
output the response to the query by the user.

2 . The system of claim 1 , wherein the processing circuitry is configured to process, for each user of a plurality of different users, records of past queries by each user of the plurality of users to generate a different knowledge graph corresponding to each user of the plurality of users.

3 . The system of claim 1 , wherein the knowledge graph comprises a chain-of-thought knowledge graph (CoT-KG) corresponding to the user, the CoT-KG modeling a chain of thought of the user.

4 . The system of claim 1 , wherein each of the edges of the knowledge graph comprises a weight that represents a probability of a co-occurrence, within a single query of the past queries by the user, of two topics represented by two nodes of the nodes joined by the edge.

5 . The system of claim 1 , wherein, to generate the knowledge graph corresponding to the user, the processing circuitry is further configured to process records of past queries by the user and corresponding past responses by the machine learning model.

6 . The system of claim 5 , wherein, to generate the knowledge graph corresponding to the user, the processing circuitry is further configured to:
for each pair of the past queries by the user and the corresponding past responses by the machine learning model, generate a chain of topics identified to be present within the pair and linked in sequential order; for each topic of the chain of topics for each pair, generate a node representing the topic within the knowledge graph; and for each two sequential topics of the chain of topics for each pair, generate an edge between two corresponding nodes of the knowledge graph.

7 . The system of claim 1 , wherein, to generate the knowledge graph corresponding to the user, the processing circuitry is further configured to process the records of past queries by the user, a profile of the user, and one or more topics of interest associated with the user.

8 . The system of claim 1 , wherein the past queries by the user comprise multimodal queries including two or more of text queries, audio queries, or video queries.

9 . The system of claim 1 , wherein the machine learning model comprises a large language model (LLM).

10 . The system of claim 1 , wherein the processing circuitry is further configured to:
receive user feedback for the response; update the knowledge graph based at least in part on the user feedback; determine, from a second query by the user, a second topic present within the second query; determine, based at least in part on the second topic present within the second query and the updated knowledge graph corresponding to the user, a second goal query, the second goal query comprising a second goal topic; provide, to the machine learning model, the second query by the user to generate, by the machine learning model, a second response to the second query by the user, wherein the second goal topic constrains the machine learning model to include the second goal topic of the second goal query within the second response; and output the second response to the second query by the user.

11 . The system of claim 10 ,
wherein the user feedback comprises a score for the query and the response, an indication of a topic present within a preceding query by the user, and an indication of the topic present within the query, and wherein to update the knowledge graph, the processing circuitry is configured to update, based at least in part on the score, a weight of an edge of the edges, the edge joining two nodes of the nodes representing the topic present within the preceding query and the topic present within the query.

12 . A method comprising:
processing, by processing circuitry, records of past queries by a user to generate a knowledge graph corresponding to the user, the knowledge graph comprising nodes and edges, each of the nodes representing a topic of topics present within the past queries by the user, and each of the edges representing a co-occurrence between the topics present within the past queries by the user; determining, by the processing circuitry, from a query by the user, a topic present within the query; determining, by the processing circuitry and based at least in part on the topic present within the query and the knowledge graph corresponding to the user, a goal query, the goal query comprising a goal topic; providing, by the processing circuitry and to a machine learning model, the query by the user to generate, by the machine learning model, a response to the query by the user, wherein the goal topic constrains the machine learning model to include the goal topic of the goal query within the response; and outputting, by the processing circuitry, the response to the query by the user.

13 . The method of claim 12 , wherein further comprising processing, by the processing circuitry and for each user of a plurality of different users, records of past queries by each user of the plurality of users to generate a different knowledge graph corresponding to each user of the plurality of users.

14 . The method of claim 12 , wherein each of the edges of the knowledge graph comprises a weight that represents a probability of a co-occurrence, within a single query of the past queries by the user, of two topics represented by two nodes of the nodes joined by the edge.

15 . The method of claim 12 , wherein generating the knowledge graph corresponding to the user comprises processing records of past queries by the user and corresponding past responses by the machine learning model.

16 . The method of claim 15 , wherein generating the knowledge graph corresponding to the user comprises:
for each pair of the past queries by the user and the corresponding past responses by the machine learning model, generating a chain of topics identified to be present within the pair and linked in sequential order; for each topic of the chain of topics for each pair, generating a node representing the topic within the knowledge graph; and for each two sequential topics of the chain of topics for each pair, generating an edge between two corresponding nodes of the knowledge graph.

17 . The method of claim 12 , wherein generating the knowledge graph corresponding to the user comprises processing the records of past queries by the user, a profile of the user, and one or more topics of interest associated with the user.

18 . The method of claim 12 , wherein the past queries by the user comprise multimodal queries including two or more of

Assignees

• Equinix Inc

Inventors

• Shastri Rajeev
• Ratnakar Shivam
• Nayak Soumya Ranjan

Classifications

• G06N5/04

  Inference or reasoning models · CPC title

• G06F40/35

  Discourse or dialogue representation · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title

• G06N3/006

  based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] · CPC title