Machine Learning Video Scoring for Low-Latency Content Selection

Solution Overview

Problem

Websites face challenges in displaying a large number of content items without causing network latency or memory constraints, leading to reduced interaction rates due to limited display capabilities.

Innovation Solution

A neural network is used to select a subset of content items likely to draw the target audience's attention, ensuring effective interaction without overwhelming network or memory resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a large number of content items are displayed on a webpage, then the target audience interaction rate increases, but network latency increases and memory constraints are exceeded

Engineering Contradiction:

Improvetarget audience interaction rateVSAvoidnetwork latency

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of content items using machine learning models before they are displayed. Interaction probability scores are pre-calculated for each content item based on target audience characteristics, allowing the system to pre-select the most promising content items without requiring real-time computation during page loading, thus reducing network latency while maintaining high interaction rates

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system extracts and displays only the most relevant content items from a larger pool. By using interaction probability scores to filter and select only the top-ranking content items that are most likely to engage the target audience, the system reduces the total number of items displayed, thereby decreasing memory usage and network bandwidth requirements while preserving interaction effectiveness

Inventive Principle:

Principle #2Taking out (Extraction)

2Productivity

If a large number of content items are displayed on a webpage, then the target audience interaction rate increases, but memory constraints are exceeded

Engineering Contradiction:

Improvetarget audience interaction rateVSAvoidmemory usage

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The system extracts and displays only the most relevant content items from a larger pool. By using interaction probability scores to filter and select only the top-ranking content items that are most likely to engage the target audience, the system reduces the total number of items displayed, thereby decreasing memory usage and network bandwidth requirements while preserving interaction effectiveness

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system changes the parameter of content item selection from displaying all available items to displaying only items above a certain interaction probability threshold. This parameter-based filtering allows dynamic adjustment of the number of displayed items based on available memory resources, ensuring that memory constraints are respected while maximizing interaction rates within those constraints

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250316057A1Systems and methods implementing a machine learning architecture for video processing

Publication Date: 2025.10.09 VIZIT LABS INC

AI summary

The present disclosure describes a method comprising receiving a video; segmenting the video into a plurality of segments, each of the plurality of segments comprising a plurality of images; executing one or more machine learning models using the plurality of segments to generate a segment score for each of the plurality of segments, the segment score for a segment indicating a likelihood that a user will interact with the segment; generating a video performance score for the video as a function of the segment scores for the plurality of segments; and generating a record comprising the video performance score for the video and an identification of the video.