Adaptive Conversation State Management for Speech Error Correction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conversational interfaces struggle with managing conversation states and filters, especially in speech recognition environments with errors, leading to inefficient and user-unfriendly information retrieval processes.

Innovation Solution

A method for adaptive conversation state management that dynamically applies filtering operators, recognizes conversation thread boundaries, and corrects user input errors by allowing users to vocalize filtering operations, using a relationship graph to manage and update the conversation state space.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If filtering operators are applied dynamically in conversational interfaces, then information retrieval precision is improved, but conversation state management complexity increases

Engineering Contradiction:

Improveinformation retrieval precisionVSAvoidconversation state management complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent implements dynamic filtering operators that adapt to conversation context in real-time. The system modifies the conversation state space by expanding or pruning states based on detected filtering operators, allowing the retrieval precision to improve dynamically without requiring a static, overly complex state management structure from the beginning

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The conversation state space is segmented into manageable components that can be independently expanded or pruned. By dividing the state management into discrete filterable units, the system handles complexity in modular fashion while maintaining high retrieval precision through selective state exploration

Inventive Principle:

Principle #1Segmentation

2Quantity of substance

If the conversation state space is expanded to handle multiple filters, then retrieval completeness is improved, but processing time increases

Engineering Contradiction:

Improveretrieval completenessVSAvoidprocessing time

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of filtering operators to predict which conversation states are likely to be relevant. By pre-processing and prioritizing state expansions based on filter characteristics, the system ensures comprehensive retrieval across multiple filters while reducing unnecessary processing time for unlikely states

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies partial expansion of the conversation state space by focusing computational resources on the most promising filter combinations. Rather than exhaustively exploring all possible states, the system performs targeted expansions that achieve sufficient retrieval completeness with reduced processing overhead

Inventive Principle:

Principle #16Partial or excessive action

3Ease of operation

If speech recognition is used for natural conversation, then ease of operation is improved, but input accuracy deteriorates due to recognition errors

Engineering Contradiction:

Improveease of operationVSAvoidinput accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system implements feedback mechanisms where filtering operators are inferred from conversational context and used to correct speech recognition errors. The conversation state provides continuous feedback that helps disambiguate recognized text, allowing the system to maintain ease of spoken operation while improving input accuracy through contextual correction

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The conversational interface performs self-correction by using the conversation state and filtering operators to automatically remedy speech recognition errors. The system serves itself by detecting inconsistencies in recognized input and correcting them based on the established conversation context, eliminating the need for manual intervention while maintaining accuracy

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS12436983B2Method for adaptive conversation state management with filtering operators applied dynamically as part of a conversational interface

Publication Date: 2025.10.07 ADEIA GUIDES INC

AI summary

A system and method of processing a search request is provided. Identification of a desired content item is based on comparing a topic of the search request to previous user input. The method includes providing access to a set of content items with metadata that describes the corresponding content items and providing information about previous searches. The method further includes receiving a present input from the user and determining a relatedness measure between the information about the previous searches and an element of the present input. If the relatedness measure is high, the method also includes selecting a subset of content items based on comparing the present input and information about the previous searches with the metadata that describes the subset of content items. Otherwise, the method includes selecting a subset of content items based on comparing the present input with the metadata that describes the subset of content items.