Secure Global Model Computation Apparatus for Federated Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In federated learning, variations in the timing of local model completion due to differences in training data and computing capability lead to inefficient processing and potential deadlocks in secure computation systems.
Innovation Solution
A secure global model computation device with N secure computation devices, equipped with a processing vacancy checking unit, transmission/reception unit, and parameter share registration unit, to efficiently receive and process local models by checking for processing vacancies and managing parameter shares.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If secure computation is used to protect training data, then data security is improved, but processing efficiency deteriorates due to variations in local model completion timing
Solution Approach 1:
The system performs preliminary actions by having local model learning devices check processing vacancy status before transmitting local models, and by having secure global model computation devices prepare reception in advance. This preliminary coordination prevents idle waiting time and ensures that devices are ready to process models as soon as they become available, thereby improving processing efficiency while maintaining secure computation protocols
Solution Approach 2:
The system implements feedback mechanisms where secure global model computation devices transmit processing vacancy status information back to local model learning devices. This feedback allows local devices to understand the current state of the secure computation system and adjust their transmission timing accordingly, eliminating unnecessary waiting and improving overall processing efficiency without compromising data security
2Reliability
If secure computation devices process models in strict sequence, then data security is maintained, but processing efficiency deteriorates due to idle waiting time
Solution Approach 1:
The system transitions from static sequential processing to dynamic parallel processing where multiple local models can be processed simultaneously by different secure global model computation devices. The processing vacancy checking mechanism dynamically coordinates this parallel processing while maintaining security protocols, thereby reducing idle waiting time and improving efficiency without compromising computation security
3Stability of the object's composition
If the secure computation system waits for all devices to be ready, then processing consistency is improved, but productivity deteriorates due to extended waiting periods
Solution Approach 1:
The system performs preliminary checks of processing vacancy status before models are transmitted, ensuring that secure global model computation devices are ready to receive and process models. This preliminary action maintains processing consistency by avoiding interruptions while minimizing waiting periods, thereby improving model reception efficiency without sacrificing processing consistency
Data Source
AI summary
Provided is a technique for efficiently receiving a local model required for computation of a global model in federated learning. A secure global model computation device in a federated learning system including M local model learning devices for learning local models using training data and a secure global model computation system composed of N secure global model computation devices for performing secure computation of a global model from M local models includes a processing vacancy checking unit configured to check whether or not the N secure global model computation devices have processing vacancy when a processing vacancy check request transmitted by one of the M local model learning devices is received and to transmit a local model transmission instruction to the local model learning device and enter a reception waiting state when all of the N secure global model computation devices have processing vacancy.