Request Prediction Using Preliminary Data Collection for Real-Time AI

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing telecommunication networks face inefficiencies in data provision and collection, particularly for real-time AI/ML tasks, due to delays in training data collection and network overload when data is collected continuously.

Innovation Solution

Implementing a requests prediction apparatus that generates information on predicted data service requests using AI/ML algorithms, allowing proactive data collection based on past requests to prepare resources ahead of time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data collection is performed continuously to ensure data availability, then data availability for real-time tasks is improved, but network overload increases

Engineering Contradiction:

Improvedata availabilityVSAvoidnetwork overload

Core Design Contradiction:

ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The system performs preliminary actions by proactively collecting training data and training AI/ML models before actual service requests arrive. The requests prediction apparatus predicts future data service requests and triggers data collection and model training in advance, ensuring data and models are ready when needed without requiring continuous data collection, thus avoiding network overload while maintaining data availability.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If data collection is performed only when needed to reduce network overhead, then network resource usage is optimized, but time delays are introduced for real-time tasks

Engineering Contradiction:

Improvenetwork resource efficiencyVSAvoiddata collection delay

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system implements feedback mechanisms where the requests prediction apparatus continuously monitors past data service requests, analyzes patterns, and uses this feedback to predict future requests. This feedback loop enables the system to proactively trigger data collection and model training before requests arrive, optimizing network resource usage by collecting data only when prediction indicates future need, while eliminating time delays by ensuring data readiness in advance.

Inventive Principle:

Principle #23Feedback

3Measurement precision

If AI/ML models are trained with large amounts of data to improve accuracy, then model performance is improved, but training time increases

Engineering Contradiction:

Improvemodel accuracyVSAvoidtraining time

Core Design Contradiction:

Measurement precisionVSDuration of action of moving object

Solution Approach 1:

The system performs preliminary model training actions by proactively collecting training data and training AI/ML models before actual service requests arrive. The requests prediction apparatus predicts future data service requests and triggers data collection and model training in advance, allowing sufficient time to gather large amounts of training data and complete model training without affecting real-time service delivery, thus achieving both high model accuracy and acceptable training time.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12542837B2Devices and methods for requests prediction

Publication Date: 2026.02.03 NOKIA TECHNOLOGIES OY

AI summary

A requests prediction apparatus (103) comprising means for: —generating information related to one or more predicted data service requests on the basis of information related to one or more past data service requests; —sending information related to at least one of the predicted data service requests to a requests prediction service client (102-i).