Deep Question Answering System Key Attribute Predictor Filtering

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Deep question answering systems incur significant processing overhead while searching and processing supporting evidence for candidate answers, which can be optimized to improve overall performance.

Innovation Solution

Implementing a deep question answering system that uses key attribute predictors to exclude inappropriate candidate answers from supporting evidence processing, thereby conserving resources and reducing unnecessary workload.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the system processes supporting evidence for all candidate answers, then the completeness of answer validation is improved, but the processing overhead and time consumption increase significantly

Engineering Contradiction:

Improveanswer validation completenessVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary filtering of candidate answers using key attribute predictors before the supporting evidence processing stage. By evaluating candidate answers against predefined key attributes early in the pipeline, the system identifies and excludes invalid candidates before investing resources in comprehensive evidence validation, thus reducing overall processing time while maintaining validation quality for promising candidates

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Instead of applying full supporting evidence processing to all candidate answers, the system applies partial processing only to candidates that pass the key attribute filter. This selective approach performs sufficient validation on promising candidates while completely skipping validation on clearly invalid ones, optimizing the balance between validation completeness and processing efficiency

Inventive Principle:

Principle #16Partial or excessive action

2Adaptability or versatility

If the system generates and processes a large set of candidate answers, then the coverage of potential answers is improved, but the processing overhead increases significantly

Engineering Contradiction:

Improveanswer coverageVSAvoidprocessing efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The system segments the candidate answer processing into two distinct stages: a rapid filtering stage using key attribute predictors and a detailed validation stage using supporting evidence. This segmentation allows the system to handle large sets of candidate answers efficiently by quickly dividing them into promising and invalid groups, then applying comprehensive processing only to the promising subset

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The key attribute evaluation is performed as a preliminary action before supporting evidence processing. This preliminary filter rapidly assesses candidate answers against essential attributes, enabling the system to maintain broad answer coverage while eliminating obviously invalid candidates before they consume resources in the more computationally intensive evidence processing stage

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If the system processes supporting evidence for all candidate answers, then the accuracy of answer validation is improved, but the resource consumption increases

Engineering Contradiction:

Improvevalidation accuracyVSAvoidcomputational resources

Core Design Contradiction:

Measurement precisionVSLoss of energy

Solution Approach 1:

The system applies partial supporting evidence processing only to candidate answers that satisfy key attribute criteria. By performing validation accuracy assessment selectively on promising candidates rather than universally on all candidates, the system maintains high validation accuracy for relevant answers while significantly reducing overall computational resource consumption

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system applies different processing qualities to different candidate answers based on their key attribute evaluation. Promising candidates receive full supporting evidence processing with high validation accuracy, while invalid candidates receive minimal or no processing. This local differentiation of processing quality optimizes resource allocation while maintaining accuracy where it matters most

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS9652717B2Avoidance of supporting evidence processing based on key attribute predictors

Publication Date: 2017.05.16 AIRBNB INC

AI summary

Systems, methods, and computer program products to perform an operation comprising receiving a case, generating a set of candidate answers for the case, excluding a first candidate answer from the set of candidate answers upon determining that a first attribute in the case precludes returning the first candidate answer as a valid response to the case, and processing supporting evidence for the remaining candidate answers in the set of candidate answers.