Forbidden Content Variant Detection via Query Log Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content filtering systems face challenges in identifying and preventing the distribution of variants of forbidden terms over networks, as they require substantial processing resources and storage to maintain and compare against large sets of potential variants, leading to inefficiencies and increased computing demands.
Innovation Solution
A system that uses query logs and semantic networks to identify and classify variants of forbidden terms by recognizing terms that cause search engines to expand queries or correct spellings, allowing for the selection of high-scoring candidate variants for inclusion in an index, thereby reducing the number of variants that need to be maintained and processed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing content filtering systems maintain and compare against large sets of potential variants of forbidden terms, then they can identify more variants, but they require substantial processing resources and storage
Solution Approach 1:
The system segments the variant detection process into multiple stages: first identifying candidate variants from query logs, then scoring and ranking them, and finally selecting only the high-scoring variants for inclusion in the forbidden terms index. This segmentation allows the system to handle large volumes of data without processing all potential variants equally, reducing overall computational burden while maintaining detection accuracy.
Solution Approach 2:
The system performs partial action by not maintaining or comparing against all possible variants of forbidden terms. Instead, it identifies and maintains only the most relevant variants based on scoring criteria derived from query log analysis. This partial approach focuses computational resources on the most likely variants that users actually employ, reducing storage and processing requirements while preserving effective filtering capability.
2Reliability
If existing content filtering systems maintain large sets of potential variants, then they can prevent more forbidden content, but they increase storage requirements
Solution Approach 1:
The system performs preliminary action by analyzing query logs in advance to identify and score candidate variants before they need to be used for filtering. This preliminary identification and scoring process allows the system to pre-select only the most relevant variants for inclusion in the forbidden terms index, reducing storage requirements while ensuring that the most commonly used variants are captured for effective filtering.
Solution Approach 2:
The system uses query logs, which are naturally generated by the search engine's operation, to automatically identify and classify variants of forbidden terms. This self-service approach leverages existing data infrastructure to populate and update the forbidden terms index without requiring external input or manual curation, reducing storage overhead by only including variants that have actually appeared in user queries.
3Adaptability or versatility
If the system identifies and maintains more variant terms, then it can block more forbidden content, but it increases the complexity of the filtering system
Solution Approach 1:
The system implements dynamics by making the forbidden terms index adaptive and evolving rather than static. It continuously analyzes query logs to identify new variants, scores them according to relevance criteria, and automatically updates the index by adding high-scoring variants and removing low-scoring ones. This dynamic approach allows the system to adapt to emerging variants without manual intervention, maintaining high variant coverage while managing complexity through automated processes.
Solution Approach 2:
The system applies parameter changes by using scoring thresholds and frequency criteria to dynamically adjust which variants are included in the forbidden terms index. By changing parameters such as minimum frequency thresholds and scoring cutoffs, the system can flexibly control the size and composition of the variant set, balancing adaptability with manageable complexity based on operational requirements.
4Reliability
If the system processes all potential variants of forbidden terms, then it achieves comprehensive filtering, but it reduces processing speed
Solution Approach 1:
The system extracts only the essential and most relevant variants from the pool of all potential variants by scoring them based on their frequency and relevance in query logs. This extraction process removes unnecessary variants from consideration, allowing the system to maintain comprehensive filtering for the most important cases while significantly reducing the computational workload and improving processing speed.
Data Source
AI summary
The subject matter of this specification generally relates to preventing the distribution of forbidden network content. In one aspect, a system includes a front-end server that receives content for distribution over a data communication network. The back-end server identifies, in the query log, a set of received queries for which a given forbidden term was used to identify a search result in response to the received query even though the given forbidden term was not included in queries included in the set of received queries. The back-end server classifies, as variants of the given forbidden term, a term from one or more queries in the set of received queries that caused a search engine to use the given forbidden term to identify one or more search results in response to the one or more queries and prevents distribution of content that includes a variant.