Communication Timing Coordination for Real-Time Distributed AI Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of improving real-time performance in data transmission between devices, particularly in AI/ML applications, where training data collection and model training are distributed across devices, is not adequately addressed by existing technologies.
Innovation Solution
A communication method and apparatus that control the generation and sending moments of data by using reference and predefined time intervals, assistance information, and association relationships to optimize data transmission, ensuring data freshness and resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If training data collection and AI model training are distributed across multiple devices, then system versatility and AI capability are improved, but data transmission real-time performance deteriorates
Solution Approach 1:
The patent applies preliminary action by having the second apparatus pre-determine and indicate the first moment (data generation moment) and second moment (data sending moment) before actual data transmission occurs. This advance scheduling allows the first apparatus to prepare and transmit data at the precisely timed moments, reducing transmission delays while maintaining distributed AI training capabilities across multiple devices.
2Measurement precision
If separate information is used to indicate generation moment and sending moment, then transmission precision is improved, but information overhead increases
Solution Approach 1:
The patent merges the indication of generation moment and sending moment into a single first information message transmitted from the second apparatus to the first apparatus. This consolidation reduces the number of separate signaling messages required, lowering information overhead while still providing precise timing control for data generation and transmission through the combined moment indications.
Data Source
AI summary
A method includes: a second apparatus sends first information to a first apparatus, where the first information is used to determine a first moment at which first data is generated and a second moment at which the first data is sent. The first apparatus sends, at the second moment, the first data.