Virtual Assistant Dialog History for Context-Aware Queries

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current virtual assistants are limited in their ability to handle conversations across multiple vertical domains and require extensive training in linguistics or artificial intelligence, making it challenging to develop systems that can understand and respond to user queries in a human-like manner, especially when dealing with incomplete or ambiguous queries.

Innovation Solution

The development of a system that extracts, stores, and retrieves dialog context information using a framework with a data structure called dialog history, allowing virtual assistants to remember relevant information from recent exchanges and use it to interpret user queries, even across different domains, without requiring deep training in linguistics or AI.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If syntax-based semantics with context-free grammar is used to process natural language queries, then syntactic structure recognition is improved, but semantic understanding and context awareness deteriorate

Engineering Contradiction:

Improvesyntactic structure recognitionVSAvoidsemantic understanding

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The system segments natural language processing into distinct layers: syntax-based semantic parsing for structured query understanding, and dialog history analysis for context retrieval. This segmentation allows each layer to specialize - the parser handles grammatical structure while the dialog history mechanism handles semantic context, resolving the contradiction between syntactic precision and semantic understanding.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system introduces dialog history as an intermediary data structure that bridges syntactic parsing and semantic understanding. The parser generates structured interpretations from syntax, which are then enriched by retrieving relevant context from dialog history, allowing semantic understanding to be built upon syntactic analysis without being limited by it.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If virtual assistants are designed to handle multiple vertical domains, then adaptability is improved, but system complexity and training requirements worsen

Engineering Contradiction:

Improvemulti-domain capabilityVSAvoidtraining requirements

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system implements a universal dialog history data structure and retrieval mechanism that functions across all vertical domains without domain-specific customization. The same core infrastructure handles context extraction, storage, and retrieval for different domains, reducing complexity while maintaining multi-domain adaptability through a single generalized system.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system copies relevant context information from dialog history into the current query interpretation process. Rather than requiring domain-specific training for each vertical, the system universally copies and applies relevant historical context to any domain query, simplifying the approach while maintaining adaptability across diverse domains.

Inventive Principle:

Principle #26Copying

3Measurement precision

If complete context is stored for every dialog exchange, then context retrieval accuracy is improved, but memory usage and processing time worsen

Engineering Contradiction:

Improvecontext retrieval accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system extracts only the most relevant context information from dialog history based on similarity matching with the current query, rather than processing or storing complete context. This extraction approach retrieves sufficient contextual information for accurate understanding while minimizing memory usage and processing time by focusing only on pertinent details.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system implements partial context retrieval by matching query patterns against dialog history and extracting only the degree of context needed for interpretation. Rather than retrieving complete historical context, the system performs partial action - retrieving just enough context to resolve the current query's meaning, balancing accuracy with efficiency.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10418032B1System and methods for a virtual assistant to manage and use context in a natural language dialog

Publication Date: 2019.09.17 SOUNDHOUND AI IP LLC

AI summary

A dialog with a conversational virtual assistant includes a sequence of user queries and systems responses. Queries are received and interpreted by a natural language understanding system. Dialog context information gathered from user queries and system responses is stored in a layered context data structure. Incomplete queries, which do not have sufficient information to result in an actionable interpretation, become actionable with use of context data. The system recognizes the need to access context data, and retrieves from context layers information required to transform the query into an executable one. The system may then act on the query and provide an appropriate response to the user. Context data buffers forget information, perhaps selectively, with the passage of time, and after a sufficient number and type of intervening queries.