Adaptive Transaction Model Orchestration for Privacy-Preserving Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing transaction processing systems face challenges in efficiently identifying potentially anomalous activities without exposing transaction data and in adapting to changing transaction patterns, often leading to false positives and inefficiencies in manual processing.
Innovation Solution
A system that orchestrates iterative updates to machine learning models deployed across multiple end-user devices, applying model parameters and weights while keeping transaction data secure, and dynamically adjusts to changing transaction patterns using peer group classifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If transaction data is shared across multiple end-user devices for model training and improvement, then model accuracy and adaptability improve, but data security and privacy protection deteriorate
Solution Approach 1:
The patent extracts only the necessary model parameters and weights from the transaction data processing, separating the learnable components from the sensitive transaction data itself. This allows model improvement without exposing the actual transaction data across end-user devices
Solution Approach 2:
The system introduces an intermediary orchestration mechanism that coordinates model updates between end-user devices. This intermediary manages the distribution of model parameters and weights, enabling collaborative model improvement while maintaining data isolation and security boundaries
2Productivity
If manual processing methods are used to identify anomalous activities, then system complexity remains low, but processing efficiency and accuracy deteriorate
Solution Approach 1:
The system implements self-service through automated machine learning models that independently identify and adapt to anomalous transaction patterns. The models automatically process transaction data, learn from patterns, and generate alerts without requiring manual review of each transaction, significantly improving processing efficiency
Solution Approach 2:
The patent utilizes parameter changes in machine learning models to dynamically adapt to changing transaction patterns. By adjusting model parameters and weights based on learned patterns, the system achieves high processing efficiency and accuracy while maintaining manageable complexity through automated adaptation
3Adaptability or versatility
If static transaction patterns are used for model training, then model stability is maintained, but adaptability to changing transaction patterns deteriorates
Solution Approach 1:
The system implements dynamics by enabling machine learning models to continuously adapt their parameters and weights based on changing transaction patterns. The models evolve over time to reflect current transaction behaviors while maintaining structural stability through controlled update mechanisms orchestrated across multiple end-user devices
4Productivity
If multiple end-user devices independently process transactions without orchestration, then system autonomy is maintained, but model refinement and knowledge sharing deteriorate
Solution Approach 1:
The system implements feedback mechanisms where each end-user device's model performance and learned patterns provide input for orchestrating updates to other devices. This feedback loop enables continuous model refinement across the distributed system, with each device benefiting from collective learning while maintaining operational autonomy
Data Source
AI summary
Systems and techniques are described for orchestrating iterative updates to machine learning models (e.g., transaction models) deployed to multiple end-user devices. In some implementations, output data generated by a first transaction model deployed at a first end-user device is obtained. The first transaction model is trained to apply a set of evidence factors to identify potentially anomalous activity associated with a first target entity. An adjustment for a second transaction model deployed at a second end-user device is determined. The second transaction model is trained to apply the set of evidence factors to identify potentially anomalous activity associated with a second target entity determined to be similar to the first target entity. A model update for the second transaction model is generated. The model update specifies a change to the second transaction model. The model update is provided for output to the second end-user device.