Personalized Related Queries Using Quick Answer Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current internet search engines fail to provide personalized query suggestions alongside quick answers, limiting the user's search experience by not offering relevant queries submitted by the user or other users that are related to the current query.
Innovation Solution
Implement a system that analyzes user query history and identifies quick answer segments using machine learning classifiers, retrieving and ranking personalized query suggestions based on user interactions and segment matching, to present them alongside quick answers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the search engine provides only quick answers without personalized query suggestions, then the search results are simple and fast to generate, but the user engagement and search experience are limited
Solution Approach 1:
The patent segments the search result page into distinct functional areas: quick answer section and personalized query suggestions section. This segmentation allows the system to provide comprehensive functionality while maintaining clear organization and user understanding, resolving the contradiction between enhanced search experience and perceived system complexity
Solution Approach 2:
The system performs preliminary actions by pre-processing user query history, pre-generating query suggestions, and pre-classifying queries into segments before the user actually searches. This preparation work happens in the background, so when the user searches, the personalized suggestions are already ready to display, making the enhanced experience feel instant rather than complex
2Loss of information
If the search engine retrieves and displays personalized query suggestions from user history, then the relevance and personalization improve, but the time and computational resources required increase
Solution Approach 1:
The patent applies local quality by retrieving and displaying only a subset of query suggestions that are most relevant to the current search context, rather than processing and displaying the entire user query history. The system selectively presents personalized suggestions based on the current query segment and user preferences, reducing processing time while maintaining high relevance
Solution Approach 2:
The system dynamically changes parameters such as the number of suggestions displayed, the time window for query history retrieval, and the relevance thresholds based on the current search context and user behavior patterns. This adaptive parameter adjustment optimizes the balance between query relevance and processing time for each specific search interaction
3Measurement precision
If the search engine classifies queries into segments using machine learning, then the accuracy of personalized suggestions improves, but the computational complexity and processing time increase
Solution Approach 1:
The patent implements partial action by using a simplified classification approach for certain query types while applying more sophisticated machine learning only when necessary. For common query patterns, the system uses rule-based or lightweight classification, reserving complex ML models for ambiguous or important queries, thus achieving good accuracy without excessive computational complexity
Solution Approach 2:
The system creates simplified copies or representations of complex query segments for classification purposes. Instead of analyzing entire complex queries with full ML models, the system extracts key features and creates simplified query representations that can be classified more efficiently while maintaining sufficient accuracy for generating relevant suggestions
Data Source
AI summary
Mechanisms to personalize quick answers to a user are described. A query history is assembled from queries previously submitted by users. Each entry in the query history comprises a query, an associated quick answer, and the userID associated with the user that submitted the query. A query database is created by submitting the query of each query entry in the query history to a trained machine learning classifier to classify the query with an associated quick answer segment. The quick answer segment is combined with the other information in the query history to create an entry into a query database. When a current query is received from a user, the query database is searched and prior queries with the same quick answer segment from the same user are extracted. A subset of the resultant queries are combined with a subset of queries from other users to personalize the search results.