Scoring Candidate Answer Passages Using Query-Dependent Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Search engines face challenges in providing relevant, explanatory answers to question queries, as existing systems often rely on scoring individual resources rather than generating and scoring candidate answer passages that combine structured and unstructured content effectively.
Innovation Solution
The method involves generating candidate answer passages from both structured and unstructured content, using query-dependent and query-independent signals to score these passages, and selecting the most relevant ones to provide as answer responses, thereby enhancing the quality and relevance of answer boxes in search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If candidate answer passages are generated from both structured and unstructured content, then the quality and relevance of answer responses is improved, but the device complexity and processing requirements increase
Solution Approach 1:
The patent segments the answer generation process into distinct components: generating candidate passages from structured content, generating candidate passages from unstructured content, scoring each candidate using query-dependent and query-independent signals, and selecting the best candidate. This segmentation allows each component to be optimized independently while maintaining overall system quality.
Solution Approach 2:
The patent performs preliminary actions by pre-computing query-independent signals and organizing structured content into passage units before query processing. This allows the system to avoid redundant computations during query time, reducing processing requirements while maintaining answer quality.
2Reliability
If candidate answer passages are generated from both structured and unstructured content, then the relevance of answer responses is improved, but the processing time increases
Solution Approach 1:
The patent applies partial action by focusing computational resources on the most promising candidate passages. Instead of exhaustively processing all possible passages, the system uses query-dependent signals to identify and score only the most relevant candidates, reducing processing time while maintaining relevance.
Solution Approach 2:
The patent changes parameters by dynamically adjusting the weighting of query-dependent versus query-independent signals based on the specific query and context. This allows the system to optimize processing time for different query types while maintaining answer relevance.
3Productivity
If query-dependent signals are weighted based on the set of most relevant resources, then the scoring analysis efficiency is improved, but the measurement precision requirements increase
Solution Approach 1:
The patent implements feedback mechanisms where the scoring system continuously refines query-dependent signal weights based on the relevance of retrieved resources. This feedback loop allows the system to improve scoring efficiency by learning from previous queries while maintaining accuracy through iterative optimization.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for scoring candidate answer passages. In one aspect, a method includes receiving a query determined to be a question query that seeks an answer response and data identifying resources determined to be responsive to the query; for a subset of the resources: receiving candidate answer passages; determining, for each candidate answer passage, a query term match score that is a measure of similarity of the query terms to the candidate answer passage; determining, for each candidate answer passage, an answer term match score that is a measure of similarity of answer terms to the candidate answer passage; determining, for each candidate answer passage, a query dependent score based on the query term match score and the answer term match score; and generating an answer score that is a based on the query dependent score.