OWT Machine Learning Modeling for Secure Payload Egress Decisions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In one-way transfer (OWT) systems, the lack of feedback opportunities for data transfers across data boundaries complicates software deployment and maintenance, particularly for machine learning models, as service providers cannot access data within the OWT system, making user training and maintenance burdensome and time-consuming.
Innovation Solution
Implement sensory and response machine learning models within the OWT system to analyze data payloads, providing insights on object classes and anomalous activity, and determine egress permissions, with sensory models trained externally and response models trained internally to facilitate accurate and efficient data management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If users deploy and train machine learning models within OWT system boundaries, then data security and isolation are improved, but user burden and complexity increase due to lack of feedback opportunities
Solution Approach 1:
The machine learning model is segmented into two distinct components: a sensory model that operates externally to analyze data payloads, and a response model that operates internally within the OWT system boundaries. This segmentation allows the sensory model to provide feedback and training opportunities externally while maintaining data security isolation, thereby reducing user burden without compromising security.
Solution Approach 2:
The sensory model acts as an intermediary between external data sources and the internal response model. It receives data payloads externally, performs analysis and generates insights, and communicates with the response model through defined interfaces. This intermediary role enables feedback opportunities and training without requiring direct access to internal system boundaries, reducing user burden while maintaining security.
2Productivity
If sensory models are trained externally and response models internally, then training efficiency and feedback opportunities are improved, but system architecture complexity increases
Solution Approach 1:
The machine learning system is segmented into sensory and response models with clearly defined roles and interfaces. The sensory model handles external data analysis and training, while the response model handles internal decision-making. This segmentation enables parallel training processes and improves efficiency while managing architecture complexity through modular design.
Solution Approach 2:
The sensory model serves multiple functions: it analyzes data payloads, detects object classes, identifies anomalous activity, and provides training feedback. This multi-functionality consolidates several operations into a single external component, improving training efficiency while the modular architecture keeps overall system complexity manageable.
Data Source
AI summary
Examples of the present disclosure describe systems and methods for sensory and response modeling in OWT systems. In examples, a payload is received by a sensory machine learning (ML) model implemented within an OWT system. The sensory ML model outputs an indication associated with data within the payload, such as whether the data belongs to one or more object classes or is indicative of anomalous activity. The output of the sensory ML model is provided to a response ML model implemented within the OWT system. The response ML model outputs a determination associated with the payload, such as whether the payload is permitted to egress across a data boundary of the OWT system or the manner in which data in the payload can be used in the one or more computing environments. The payload is then processed in accordance with the determination.