Ad Relevance Scoring via Distributed Ordinal Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content-targeted ad delivery systems face inefficiencies in filling ad spots, dilute relevance information, and increase complexity due to sequential processing and equal treatment of criteria, leading to wasted opportunities and resource overload.
Innovation Solution
Implementing a scoring system that ranks ad relevance using a price parameter and ordinal ranking of criteria, allowing for distributed selection and filtering of ads based on their relevance to document content, thereby optimizing ad placement and reducing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sequential processing of criteria is used to select ads, then ad relevance to document content is improved, but processing time and system complexity increase
Solution Approach 1:
The patent pre-ranks criteria by relevance to document content before ad selection. By determining an ordinal ranking of criteria in advance (prior to ad selection), the system prepares the selection hierarchy beforehand, enabling faster ad retrieval without compromising relevance during the actual selection process
Solution Approach 2:
The patent divides the ad selection process into distinct phases: criteria ranking phase and ad selection phase. By segmenting the overall process and pre-computing criterion rankings, the system reduces the computational burden during real-time ad selection, thereby reducing processing time while maintaining relevance
2Measurement precision
If all criteria are treated equally in ad selection, then system simplicity is maintained, but ad relevance and quality decrease
Solution Approach 1:
The patent applies different weights to different criteria based on their local quality or relevance to the document. By assigning ordinal rankings to criteria (where some criteria are deemed more relevant than others), the system enhances ad relevance by prioritizing matches on high-quality criteria without requiring complex weighted calculations for all criteria
3Measurement precision
If sequential ad selection process is used, then ad relevance is improved, but number of ad spots filled decreases
Solution Approach 1:
By pre-ranking all criteria by relevance before selection, the system can efficiently evaluate multiple ads against the ranked criteria hierarchy. This preliminary organization enables the system to fill more ad spots by quickly identifying relevant ads without sacrificing relevance quality
Solution Approach 2:
The patent changes the parameter of criterion evaluation from equal-weighted to ordinal-ranked. This parameter change allows the system to efficiently process multiple ad candidates against the ranked criteria, increasing productivity by filling more ad spots while maintaining high relevance through the ordinal ranking system
4Use of energy by moving object
If distributed ad selection is implemented, then resource load is reduced, but system complexity increases
Solution Approach 1:
The patent segments the ad selection workload across multiple distributed servers. Each server can independently handle criteria ranking and ad selection for its assigned document or criterion set. This segmentation reduces the processing load on any single server while the standardized ordinal ranking approach keeps architecture complexity manageable
Data Source
AI summary
Ads eligible to be served with a document (for example, because they are relevant to the document) may each be scored using a price parameter associated with the ad and an indication of relevancy of the ad to the document. The indication of relevancy of the ad to the document may be based on an ordinal ranking of a relevancy criteria of the document used to select the ad, and/or a value of a relevancy criteria of the document used to select the ad. The eligible ads may be determined by obtaining relevancy criteria for the document and selecting ads using at least some of the obtained relevancy criteria. The ads may be selected, and perhaps filtered, in a distributed manner.