Neural Network Image Selection for Webpage Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Websites face challenges in displaying a large number of content items due to browser interface size, internet bandwidth, and memory restrictions, leading to significant latency and reduced interaction rates.
Innovation Solution
A computer system uses a neural network to automatically select a small number of content items that are likely to draw the target audience's attention, by extracting images from a webpage and generating a target audience interaction benchmark, and then uploading only the identified images to the website.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large number of content items are displayed on a webpage, then the target audience interaction rate increases, but the browser interface size, internet bandwidth, and memory restrictions cause significant latency and technical failures
Solution Approach 1:
The patent segments the large set of content items into multiple smaller batches or groups. Instead of loading all content items at once, the system divides them into manageable portions that can be loaded sequentially or selectively, reducing the memory and bandwidth requirements at any given time while still providing access to a large number of content items overall.
Solution Approach 2:
The patent extracts and prioritizes the most relevant or engaging content items from the larger set based on predefined criteria or user preferences. By selecting only the most valuable content items for initial display, the system reduces the number of items loaded into the browser interface, thereby reducing latency and resource consumption while maintaining high interaction rates.
2Productivity
If a large number of content items are displayed on a webpage, then the target audience interaction rate increases, but the internet bandwidth causes significant latency
Solution Approach 1:
The patent performs preliminary actions by pre-processing and pre-selecting content items before they are displayed to the user. This includes pre-loading metadata, pre-rendering thumbnails, or pre-fetching content based on predicted user interest, so that when the user interacts with the page, the actual content is already prepared or quickly accessible, reducing perceived latency.
Solution Approach 2:
The patent implements partial loading by displaying a subset of content items initially and loading additional items as the user interacts with the page. This progressive loading approach ensures that the most important content is available immediately, while less critical content is loaded on-demand, reducing initial latency while still providing access to a large number of content items.
3Productivity
If a large number of content items are displayed on a webpage, then the target audience interaction rate increases, but the memory restrictions cause significant latency or prevent content display
Solution Approach 1:
The patent implements a nested structure where content items are organized in hierarchical levels. The browser interface displays a summary or thumbnail level first, and when users interact with specific items, more detailed content is loaded in nested layers. This allows the system to manage large amounts of content data efficiently by only loading detailed information for items that users actually interact with, reducing overall memory usage.
Solution Approach 2:
The patent applies local quality by optimizing the representation of content items based on their location or importance in the interface. High-priority or frequently accessed content items are stored in higher-quality formats with more detailed information, while less important items use compressed or summarized representations. This selective quality approach reduces overall memory requirements while maintaining the quality of critical content.
Data Source
AI summary
A method includes receiving a first image; extracting a first set of features from the first image; executing a first machine learning model using the extracted first set of features as input to generate a first image performance score for the first image; iteratively executing, using the first set of features as input for each execution, a generative machine learning model to generate a plurality of generated images; extracting a second set of features from each of the plurality of generated images; iteratively executing the first machine learning model using a different second set of features as input for each execution to generate a generated image performance score for each of the plurality of generated images; and transmitting a defined number of the plurality of generated images to a computing device, the defined number of generated images determined based on the generated image performance scores.