AI Content Push Prioritization Under Dynamic Resource Input
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content push methods result in inaccurate content selection and waste of hardware resources due to estimating virtual resource inputs in an offline state, leading to low accuracy and inefficient use of system resources.
Innovation Solution
A content push method that determines candidate mixed content comprising first-type content based on dynamic resource input and second-type content as agreed, using artificial intelligence to calculate priority push probabilities based on object attributes, dynamic resource inputs, and content features to optimize content selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If content is selected based on estimated virtual resource input amount and resource input amount in offline state, then content push process is simplified, but content push accuracy deteriorates and hardware resources are wasted
Solution Approach 1:
The patent performs preliminary sorting of content in offline state based on virtual resource input amount and resource input amount to pre-establish a content pool. When a push request arrives, the system only needs to select from this pre-sorted pool rather than sorting all content from scratch, thus maintaining operational simplicity while improving accuracy through subsequent real-time probability calculations.
Solution Approach 2:
The patent introduces dynamic priority push probability calculations that adjust content selection in real-time based on current resource input amounts and content features. This dynamic adjustment mechanism allows the system to maintain simplicity in the overall process while achieving high accuracy through adaptive, real-time decision-making for content selection.
2Productivity
If content is directly selected from pre-sorted content based on offline estimates, then processing speed is improved, but content push accuracy deteriorates
Solution Approach 1:
The system performs preliminary sorting of content in offline state based on virtual resource input amount and resource input amount to pre-establish a content pool. This preliminary action reduces the scope of real-time processing, maintaining high speed while improving accuracy.
Solution Approach 2:
The patent implements a feedback mechanism where the system calculates priority push probabilities in real-time based on current resource input amounts and content features, then uses this feedback to select optimal content. This feedback loop ensures high accuracy while maintaining processing speed by only evaluating probabilities for a reduced set of pre-sorted content.
3Device complexity
If offline estimation of virtual resource input amount is used, then system complexity is reduced, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent performs preliminary sorting of content in offline state based on virtual resource input amount and resource input amount. This preliminary action reduces the number of content items that need to be evaluated in real-time, thereby reducing hardware resource consumption during push operations while keeping the system relatively simple.
Solution Approach 2:
The system performs sorting on a partial set of content (only those meeting basic offline criteria) rather than all content. This partial action approach reduces computational overhead and hardware resource usage while maintaining sufficient accuracy for effective content push.
Data Source
AI summary
This application relates to a content push method performed by a computer device, including: determining candidate mixed content in response to a content push request from an object, the mixed content including first-type content and second-type content; determining a first priority push probability of the first-type content based on the object attribute information, the overall content feature, and the dynamic resource input feature; analyzing a priority push relationship between the content features of all pieces of second-type content, to obtain an internal priority push feature, and performing inter-type priority push relationship analysis based on the object attribute information, the internal priority push feature, and the dynamic resource input feature, to determine second priority push probabilities respectively corresponding to all pieces of second-type content; and selecting content from the mixed content for push based on the first priority push probability and the second priority push probabilities.