Chain-of-Thought Knowledge Graphs for Multi-Session Query Context

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional AI-based conversational assistants struggle to understand and remember the context of conversations spanning multiple sessions or wide-ranging topics, predict user queries accurately, and maintain and update knowledge graphs in real-time scenarios.

Innovation Solution

A conversational assistant system employs a chain-of-thought knowledge graph (CoT-KG) to model a user's thought process, integrating user-specific interactions to predict user intents and future queries, and update knowledge graphs in real-time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional AI-based conversational assistants use machine learning models to understand and respond to user inputs, then the assistant can generate human-like responses to a wide range of questions, but the assistant struggles to understand and remember the context of conversations spanning multiple sessions or wide-ranging topics

Engineering Contradiction:

Improveability to understand and remember conversation contextVSAvoidcontext memory across multiple sessions

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system performs preliminary action by proactively predicting the next user query before the user actually asks it. The knowledge graph models the user's thought process and allows the assistant to anticipate future questions, enabling the assistant to maintain context across sessions by preparing responses in advance based on predicted user needs.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses feedback by continuously monitoring user interactions and updating the knowledge graph with new information about the user's preferences, topics of interest, and question patterns. This feedback loop enables the assistant to improve its context understanding and memory capabilities over time by learning from actual user behavior.

Inventive Principle:

Principle #23Feedback

2Measurement precision

If the assistant uses knowledge graphs to enhance understanding of user queries, then the bot can better understand the context of a user's question and provide more accurate responses, but the knowledge graphs are difficult to maintain and update in real-time scenarios

Engineering Contradiction:

Improveaccuracy of understanding user contextVSAvoidmaintenance and updating of knowledge graphs

Core Design Contradiction:

Measurement precisionVSEase of manufacture

Solution Approach 1:

The system implements self-service by automatically updating the knowledge graph without requiring manual intervention. The assistant continuously monitors user interactions, extracts relevant information, and updates its own knowledge graph in real-time, making the maintenance process autonomous and eliminating the need for separate manual updates.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system ensures continuity of useful action by maintaining a continuously updated knowledge graph that reflects the user's evolving preferences and topics of interest. The knowledge graph is updated in real-time during each interaction, ensuring that the assistant always has the most current and relevant information available.

Inventive Principle:

Principle #20Continuity of useful action

3Ease of operation

If the assistant predicts the next question or topic the user is likely to ask, then the assistant can provide a seamless and engaging user experience, but accurate prediction remains a significant challenge in current technology

Engineering Contradiction:

Improveseamless and engaging user experienceVSAvoidaccuracy of next query prediction

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system performs preliminary action by modeling the user's thought process in advance and predicting future queries before they are actually asked. The knowledge graph captures the user's cognitive trajectory, allowing the assistant to anticipate and prepare responses for predicted questions, thereby creating a more seamless and engaging user experience.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses feedback by continuously monitoring user interactions and adjusting its predictions based on actual user behavior. The knowledge graph is updated with new information about the user's preferences and question patterns, enabling the assistant to improve the accuracy of its predictions over time through iterative learning from user feedback.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20260057254A1User profiling using chain-of-thought knowledge graphs for querying a machine learning system

Publication Date: 2026.02.26 EQUINIX INC

AI summary

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.