Neural Network Image Scoring for Web Search Latency
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 latency issues and reduced interaction rates.
Innovation Solution
A neural network is used to select a limited number of content items that are likely to draw the target audience's attention by generating a benchmark from extracted images and comparing new images against it, thereby reducing the number of items displayed on a webpage to avoid technical issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large number of content items are displayed on a webpage to increase interaction rates, then engagement increases, but latency and technical performance deteriorate due to browser interface size, bandwidth, and memory restrictions
Solution Approach 1:
The patent extracts and displays only the most relevant content items from a larger set, using neural network-based image scoring to identify and extract high-value content. This selective extraction maintains interaction rates while reducing the total number of displayed items, thereby decreasing latency and bandwidth consumption.
Solution Approach 2:
The system changes the parameter of content selection by introducing image quality scores generated through neural networks. Instead of displaying all content items or using simple ranking metrics, the system transforms the selection criterion to include visual appeal scores, enabling smarter curation that reduces item count while maintaining engagement.
2Productivity
If more content items are displayed to achieve high interaction rates, then engagement improves, but device memory and network bandwidth are overwhelmed
Solution Approach 1:
The system extracts only the essential high-value content items based on neural network image scoring, removing unnecessary content from the display set. This extraction process reduces data volume transmitted over the network and stored in device memory while preserving the content that drives user interaction.
Solution Approach 2:
Rather than displaying all available content or using overly aggressive filtering, the system applies partial action by selecting a optimized subset of content items. The neural network scores allow the system to display enough content to maintain engagement without exceeding bandwidth and memory constraints.
3Loss of time
If the number of displayed content items is reduced to avoid latency issues, then performance improves, but interaction rates may decrease
Solution Approach 1:
The system changes the selection parameter from quantity-based to quality-based metrics. By using neural network-generated image scores, the system identifies content with higher visual appeal and engagement potential, ensuring that fewer displayed items can achieve the same or better interaction rates.
Solution Approach 2:
The patent replaces traditional mechanical or rule-based content selection methods with a neural network-based scoring system. This substitution enables more intelligent content curation that can identify high-engagement content more accurately, allowing reduced content volume to maintain interaction rates.
Data Source
AI summary
A method includes storing at least one image performance score for each of a set of images, the set of images comprising a plurality of subsets of images, each subset corresponding with a different web page of a plurality of web pages, the at least one image performance score for an image indicating a likelihood that a user will interact with the image; determining a web page score for each of the plurality of web pages based on one or more image performance scores of the subset of images that corresponds with the web page; receiving a query comprising one or more keywords or images; selecting a set of web pages by applying a search engine machine learning model to the one or more keywords and the web page score for each of the plurality of web pages; and presenting the set of web pages at a computing device.