Media Search Engine Query Rewriting via Web Intermediary
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current media search engines struggle to provide relevant results when search queries include keywords not indexed in the database, often returning irrelevant results due to the descriptive nature of queries, such as movie plots or storyline descriptions.
Innovation Solution
A system that processes media search queries by identifying relevant web search results based on relevancy scores, extracting keyword and contextual information, assigning topic scores to media entities, and selecting media assets from a corpus to present relevant content to users, incorporating mechanisms for rewriting queries, parsing information, and determining quality scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a traditional indexing database is used for media search, then search speed is improved, but search accuracy deteriorates when queries include non-indexed keywords
Solution Approach 1:
The patent introduces web search results as an intermediary layer between the user's descriptive query and the indexed media database. When a user submits a query with non-indexed keywords (e.g., movie plot descriptions), the system first performs a web search to find relevant content, then uses the web results to identify and retrieve matching media assets from the indexed database, thereby maintaining both search speed and accuracy
Solution Approach 2:
The search process is segmented into multiple stages: (1) initial web search for queries with non-indexed keywords, (2) extraction of media entities from web results, (3) assignment of topic scores to entities, and (4) retrieval of media assets from the indexed database. This segmentation allows the system to handle different query types appropriately while maintaining overall efficiency
2Ease of operation
If the search engine accepts descriptive queries (e.g., movie plots), then user ease of operation is improved, but search reliability deteriorates due to irrelevant results
Solution Approach 1:
The system implements feedback mechanisms through topic scores and relevancy scoring. Web search results are evaluated and scored based on their relevance to the query and the media entities they contain. This feedback loop allows the system to identify the most relevant media entities and retrieve accurate results, maintaining reliability even when accepting flexible, descriptive user queries
Solution Approach 2:
Web search results serve as an intermediary that bridges the gap between user-friendly descriptive queries and the structured indexed database. The system uses this intermediary to translate natural language descriptions into precise media asset retrievals, maintaining both ease of operation and search reliability
3Quantity of substance
If the system processes all web search results, then search completeness is improved, but processing time increases
Solution Approach 1:
The system applies partial action by processing only a subset of web search results rather than all results. It uses topic scoring to identify and focus on the most relevant media entities from the web results, extracting information from only those results that contribute meaningfully to the search, thereby maintaining completeness while reducing processing time
Data Source
AI summary
Methods, systems, and media for providing a media search engine are provided. In some implementations, a method for searching for content is provided, the method comprising: receiving a media search query for media assets; identifying web search results from a corpus of web resources that are responsive to the media search query, wherein each of the web search results has an associated relevancy score; determining keyword information and contextual information from a subset of the web search results, wherein the subset of the web search results is selected from the web search results based on the associated relevancy score and wherein media entities are determined from at least a portion of the keyword information and the contextual information; assigning a topic score for the each of the media entities based on occurrence in the web search results; selecting at least one media entity from the media entities based on the topic score; identifying media assets from a corpus of media assets based at least in part on the selected media entity; and causing a subset of the media assets to be presented to a user in response to the media search query.