Machine Learning Resource Forecasting for Gaming Platforms

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Collaboration platforms, such as gaming platforms, face challenges in estimating resource usage due to time-varying demand for evaluating user reports, which can lead to insufficient or excessive resource allocation.

Innovation Solution

A machine learning model is trained using historical data of report submissions and resource availability to forecast resource requirements for future time periods, ensuring efficient resource allocation even with unknown or varying report volumes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If resource allocation is increased to handle peak report volumes, then report evaluation reliability is improved, but resource waste increases during low-demand periods

Engineering Contradiction:
Improvereport evaluation reliabilityVSAvoidresource waste
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The system dynamically adjusts resource allocation based on predicted report volumes rather than maintaining fixed capacity. The machine learning model continuously forecasts future report submission patterns and the system provisions resources accordingly, allowing the system to adapt its resource level from idle to active state based on actual demand predictions.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system performs preliminary resource provisioning based on predicted future report volumes. By using historical data and machine learning models to forecast when peak demand will occur, the system proactively allocates resources before the actual peak hits, ensuring adequate capacity is ready without over-provisioning during low-demand periods.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If resource allocation is reduced to minimize waste, then resource efficiency is improved, but report evaluation reliability deteriorates during peak demand

Engineering Contradiction:
Improveresource efficiencyVSAvoidreport evaluation reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system transitions from static resource allocation to dynamic, demand-driven resource provisioning. By continuously monitoring historical patterns and predicting future report volumes, the system can confidently reduce resources during low-demand periods while maintaining the ability to rapidly scale up when peak demand is predicted, thus avoiding both waste and reliability issues.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses feedback from historical report data to continuously refine its predictions. By analyzing past report submission patterns, resource usage, and evaluation outcomes, the machine learning model improves its forecasting accuracy, enabling more precise resource allocation decisions that balance efficiency and reliability.

Inventive Principle:
Principle #23Feedback

3Device complexity

If manual resource estimation methods are used, then system complexity is reduced, but forecasting accuracy deteriorates due to inability to handle time-varying demand

Engineering Contradiction:
Improvesystem complexityVSAvoidforecasting accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The system performs self-service through automated machine learning models that continuously learn from historical data without requiring manual intervention. The system automatically provisions, monitors, and adjusts resources based on its own predictions, eliminating the need for manual estimation while achieving high forecasting accuracy through data-driven insights.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical estimation processes with automated machine learning algorithms. Instead of relying on human analysts to interpret historical data and predict future demand, the system uses computational models that process large datasets, identify patterns, and generate precise forecasts automatically, significantly improving accuracy while reducing operational complexity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3750116B1Using machine learning to estimate or forecast resource use with time-varying demand in gaming platforms
Publication Date: 2025.04.09 ROBLOX CORP
  • EP3750116B1 patent drawingFigure 1
  • EP3750116B1 patent drawingFigure 2
  • EP3750116B1 patent drawingFigure 3

AI summary

A method and system is disclosed for training a machine learning model by generating first training input that includes a first number of reports at a first point in time. The reports are submitted by the users of the gaming platform and identify incidents where content of the gaming platform violates a policy of use associated with the gaming platform. The method and system generates second training input including a number of resources active at the first point in time. The method and system generates first target output identifies a number of resources sufficient to evaluate the target percentage of the first number of reports. The method and system provide the training data to train the machine learning model on (i) a set of training inputs comprising the first training input and the second training input, and (ii) a set of target outputs comprising the first target output.