AI Advertising Detection in Search Result Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI systems lack the ability to effectively differentiate between advertising and non-advertising web pages in search results, leading to user inefficiency in finding relevant information.
Innovation Solution
An electronic device and method that utilize a second server trained with an AI algorithm to generate additional information indicating the probability of each web page being an advertising page, allowing the device to filter and display search results based on this information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AI algorithm is used to generate additional information about advertising pages, then search result relevance is improved, but device complexity increases
Solution Approach 1:
The patent introduces a second server as an intermediary component that hosts the AI training model and processes web pages to generate additional information about advertising patterns. This mediator handles the complex AI processing tasks externally, allowing the electronic device to benefit from improved search result relevance without bearing the full burden of the computational complexity.
Solution Approach 2:
The system is divided into multiple independent components: the electronic device for user interaction, the first server for basic search functionality, and the second server for AI-based advertising detection. This segmentation allows the complex AI processing to be isolated in a dedicated server, reducing the complexity burden on individual devices while maintaining high measurement precision for search result relevance.
2Loss of information
If additional information is generated and transmitted for each web page, then information accuracy is improved, but communication data volume increases
Solution Approach 1:
The additional information generated by the AI model is not uniformly applied to all search results but is selectively generated based on the specific characteristics of each web page and the user's search context. This local quality approach ensures that information accuracy is improved only where necessary, rather than adding redundant data to all results, thereby controlling communication data volume.
Solution Approach 2:
The system generates additional advertising pattern information selectively for web pages that require it, rather than processing every single search result with the full AI model. This partial action approach maintains information accuracy for critical cases while reducing the overall communication data volume by avoiding unnecessary processing of clearly non-advertising pages.
3Measurement precision
If AI training model is used to detect advertisement patterns, then detection precision is improved, but processing time increases
Solution Approach 1:
The AI training model is pre-trained offline with extensive advertising pattern data before being deployed. This preliminary action allows the model to learn complex advertising patterns in advance, so that during actual search operations, the detection process is faster and more precise without requiring extensive real-time processing.
Solution Approach 2:
The patent uses a trained AI model that has been copied and deployed on the second server. This copied model can quickly process web pages using the knowledge it learned during offline training, providing high detection precision without requiring the electronic device to perform time-consuming training computations in real-time.
Data Source
AI summary
Provided are an electronic device and an operating method thereof. The electronic device includes a communicator, a display, and a processor configured to receive search result data based on a search keyword from a first server through the communicator, request, from a second server through the communicator, additional information indicating whether each web page included in the search result data is an advertising page, and control the display to display the search result data, based on the additional information corresponding to the each web page and received from the second server, wherein the additional information is generated by the second server by using a training model trained using an artificial intelligence (AI) algorithm, based on whether the web page includes an advertisement pattern.Also provided are an AI system using a machine learning algorithm such as deep learning and applications thereof.