Virtual Agent Context Transformer for Unconventional Query Handling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing automated query response systems are ineffective in providing useful responses to queries posed in unconventional patterns and fail to maintain context across multiple related queries, leading to suboptimal responses.

Innovation Solution

A computer-implemented method using a machine learning-based virtual dialogue system that identifies unfilled dialogue slots, generates contextualized responses and questions through transformer models, and computes slot values using a span prediction model to improve predictive responses.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If regex or similar search techniques are used for automated query response, then the system can provide responses to conforming queries, but it fails to provide responses to non-conforming queries posed in unconventional patterns

Engineering Contradiction:

Improveability to handle unconventional query patternsVSAvoidresponse accuracy for non-conforming queries

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent replaces rigid regex-based mechanical search techniques with flexible transformer-based language models that can naturally understand and respond to unconventional query patterns. The transformer model processes queries through attention mechanisms rather than pattern matching, enabling adaptation to diverse query formulations while maintaining reliable response generation.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the fundamental parameters of query processing by transitioning from fixed pattern templates to dynamic contextual understanding. The transformer model adjusts its processing based on the semantic content and structure of each query, allowing it to handle non-conforming queries by adapting to their unique characteristics rather than requiring predefined patterns.

Inventive Principle:

Principle #35Parameter changes

2Loss of information

If traditional automated query response systems are used, then the system structure remains simple, but it cannot maintain context across multiple related queries

Engineering Contradiction:

Improvecontext retention across multiple queriesVSAvoidsystem architecture complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent implements preliminary context encoding by processing and storing contextual information from previous queries before they are needed for response generation. The transformer model pre-processes dialogue history and maintains contextual representations that are readily available when subsequent queries arrive, enabling effective context retention without complex real-time processing.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system employs a nested architecture where contextual information from multiple queries is embedded within hierarchical representations. The transformer model nests previous query contexts within cumulative dialogue states, allowing each new query to access and integrate information from the entire conversation history in a structured, layered manner.

Inventive Principle:

Principle #7Nested doll (Nesting)

Data Source

PatentUS12271407B2Systems and methods for implementing a virtual agent performing context and query transformations using unsupervised machine learning models

Publication Date: 2025.04.08 KNOWBL LLC

AI summary

A computer-implemented method for improving a predictive response of a machine learning-based virtual dialogue agent includes identifying an unfilled dialogue slot associated with an active dialogue between a user and the machine learning-based virtual dialogue agent, obtaining subsequent dialogue input data based on prompting the user for the unfilled dialogue slot, and computing a value of the unfilled dialogue slot based on the obtaining of the subsequent dialog input, wherein computing the value of the unfilled dialogue slot includes computing, via a question-answering machine learning model, a slot answer inference comprising the value of the unfilled dialogue slot based on an input of a machine learning-derived query and machine learning-derived context computed for the unfilled dialogue slot.