User-Specific Ranking Model for Search Result Personalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines use generic ranking criteria that do not account for individual user preferences, leading to unsatisfactory search results for users with different search intents, despite submitting the same query.
Innovation Solution
A method and system for generating a user-specific ranking model on an electronic device, which receives resource-specific features from a search engine server, determines user interactions, and applies machine learning algorithms to create a user-centric ranking model that optimizes search result relevance based on user preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If generic ranking criteria are used to rank search results, then the search engine can process queries efficiently with simple algorithms, but the search results do not account for individual user preferences and search intents
Solution Approach 1:
The patent segments the ranking model into two distinct components: a generic ranking model that handles basic search result ranking using simple criteria, and a user-specific ranking model that captures individual user preferences and behaviors. This segmentation allows the system to maintain processing efficiency through the generic model while adding personalization through the user-specific model, thereby resolving the contradiction between efficiency and adaptability.
Solution Approach 2:
The patent introduces a user-specific ranking model as an intermediary layer between the search engine and the user. This intermediary model processes user interactions and preferences locally on the user's device, then combines its output with the generic ranking results. This intermediary approach enables personalization without requiring the search engine to process complex user data centrally, thus maintaining efficiency while improving adaptability.
2Reliability
If user interaction data is collected and processed on a search engine server to create personalized ranking models, then search result relevance can be improved, but user privacy is compromised and server-side data tracking increases
Solution Approach 1:
The patent extracts the user-specific ranking model generation and processing functionality from the search engine server and places it locally on the user's electronic device. By taking out the data processing operations from the server, the system can create personalized ranking models using user interaction data without transmitting sensitive information to the server, thereby improving search result relevance while preserving user privacy.
Solution Approach 2:
The patent implements self-service by enabling the user's device to autonomously generate and maintain the user-specific ranking model using local machine learning algorithms. The device automatically processes user interactions, updates the ranking model, and applies it to search results without requiring server-side data collection or processing. This self-service approach improves reliability through personalization while eliminating privacy concerns associated with server-side tracking.
3Adaptability or versatility
If a user-specific ranking model is generated and maintained on the user's electronic device, then user privacy is preserved and personalization is achieved, but the device complexity and computational resources required increase
Solution Approach 1:
The patent applies preliminary action by pre-training the user-specific ranking model using offline data and algorithms before deployment on the user's device. The complex model training and data processing operations are performed in advance, and the pre-trained model is then transferred to the user's device for lightweight inference and updates. This preliminary action reduces the computational burden on the user's device while maintaining personalization capability.
Solution Approach 2:
The patent uses copying by transferring the user-specific ranking model from a training environment to the user's electronic device. The model is copied in a compact, optimized form that can be efficiently stored and executed on resource-constrained devices. This copying approach enables personalization without requiring the device to perform complex training operations, thereby reducing device complexity while maintaining adaptability.
Data Source
AI summary
There is disclosed a method of generating a user-specific ranking model on an electronic device associated with a user. The method is executable on the electronic device. The method comprises: receiving, from a search engine server, via a communication network, an indication of a resource-specific feature; appreciating a user interaction with the web resource performed by the user using the electronic apparatus; based on the user interaction, determining a value parameter for the web resource; generating the user-specific ranking model on the basis of the value parameter and the resource-specific feature.