ML-Based User Experience Prediction for Online Conferencing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network technologies face challenges in predicting and ensuring user experience metrics for online conferencing, particularly in determining Service Level Agreements (SLAs) and identifying real Quality of Service (QoS) provided to applications, especially in sensitive traffic like conferencing, which is prone to noticeable delays and quality issues.
Innovation Solution
A machine learning-based approach is employed to predict user experience metrics for online conferencing by analyzing network data and user feedback, allowing endpoint nodes to switch to alternative connections based on predicted metrics, using techniques such as data collection, anonymization, mapping, and machine learning modeling to provide proactive quality assurance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional network monitoring and SLA mechanisms are used, then network performance can be monitored, but user experience metrics cannot be accurately predicted for conferencing applications
Solution Approach 1:
A machine learning model acts as an intermediary between network data and user experience metrics. The model receives network performance data as input and predicts user experience metrics (such as MOS scores) as output, enabling accurate prediction without directly measuring user experience. This intermediary processing layer resolves the contradiction by transforming complex network data into meaningful predictions.
Solution Approach 2:
The patent replaces traditional mechanical/network monitoring approaches with a data-driven machine learning system. Instead of using conventional threshold-based monitoring and manual SLA enforcement, the system uses trained ML models to predict user experience metrics, substituting mechanical monitoring with intelligent prediction capabilities.
2Reliability
If network data is collected and analyzed in real-time, then user experience can be monitored, but processing time and computational resources increase
Solution Approach 1:
The machine learning model is trained in advance on historical network data and user feedback before deployment. This preliminary training phase allows the model to learn patterns and relationships offline, so that during real-time operation, predictions can be made quickly using the pre-learned knowledge, reducing processing time while maintaining reliability.
Solution Approach 2:
The system incorporates user feedback (such as MOS scores) to continuously improve the machine learning model. This feedback mechanism allows the system to learn from actual user experiences and refine its predictions over time, enhancing reliability without requiring increased real-time processing resources.
3Ease of operation
If endpoint nodes switch connections based on predicted metrics, then user experience improves, but network control complexity increases
Solution Approach 1:
Endpoint nodes are empowered to make their own connection switching decisions based on predicted user experience metrics. Each endpoint autonomously evaluates predictions and selects optimal connections without requiring centralized network control for each decision. This self-service approach improves ease of operation while actually reducing network control complexity by distributing the decision-making process.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
In embodiments, a device in a network receives an indication of a connection between an endpoint node in the network and a conferencing service. The device retrieves network data associated with the indicated connection between the endpoint node and the conferencing service. The device uses a machine learning model to predict an experience metric for the endpoint node based on the network data associated with the indicated connection between the endpoint node and the conferencing service. The device causes the endpoint node to use a different connection to the conferencing service based on the predicted experience metric.