Distributed AI Model Execution with Secure Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI model execution methods are inefficient on constrained computing devices and vulnerable to adversarial attacks, as they do not effectively manage resource utilization and security in distributed inference tasks.
Innovation Solution
The AI model is split into input, intermediate, and output blocks, with specific encoding protocols used for secure communication and resource management across distributed computer systems, allowing efficient execution on edge devices and cloud environments while mitigating security risks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If AI model is executed on constrained computing devices, then resource utilization efficiency is improved, but security vulnerability to adversarial attacks increases
Solution Approach 1:
The AI model is divided into multiple blocks (first block, second block, third block) that are distributed across different computing devices. The first block resides on the constrained device, the second block on a remote device, and the third block back on the constrained device. This segmentation allows the model to run on resource-constrained devices while leveraging remote resources for secure intermediate processing, thus improving resource utilization without compromising security.
2Productivity
If AI model is split and distributed across multiple systems, then execution efficiency on constrained devices is improved, but system complexity increases
Solution Approach 1:
Encoding protocols act as intermediaries between the distributed blocks. The first intermediate output is encoded before transmission to the remote device, and the second intermediate output is encoded before returning to the constrained device. These encoding protocols simplify the communication interface between distributed systems, managing system complexity while enabling efficient distributed execution.
3Reliability
If encoding protocols are used for secure transmission, then security against adversarial attacks is improved, but computational overhead increases
Solution Approach 1:
Encoding is applied selectively only to intermediate outputs that are transmitted between devices, rather than encoding all data throughout the entire model execution pipeline. This partial application of encoding provides necessary security protection for transmitted data while minimizing computational overhead by avoiding redundant encoding operations on data that remains on the same device.
Data Source
AI summary
The present disclosure relates to a method for executing an artificial intelligence model, including receiving an input for execution of the AI model. An input block of the AI model can be executed by a first computer system using the received input, producing a first output. The first output can be encoded. The encoded first output can be sent to a second computer system. The second computer system can decode the encoded first output. The second computer system can execute an intermediate block of the AI model using the first output, producing a second output. The second output can be encoded. The encoded second output can be sent to the first computer system. The first computer system can decode the encoded second output. The first computer system can execute an output block of the ML model using as input the second output, producing a result output.