Abstract:Deploying DL models on mobile Apps has become ever-more popular. However, existing studies show attackers can easily reverse-engineer mobile DL models in Apps to steal intellectual property or generate effective attacks. A recent approach, Model Obfuscation, has been proposed to defend against such reverse engineering by obfuscating DL model representations, such as weights and computational graphs, without affecting model performance. These existing model obfuscation methods use static methods to obfuscate the model representation, or they use half-dynamic methods but require users to restore the model information through additional input arguments. However, these static methods or half-dynamic methods cannot provide enough protection for on-device DL models. Attackers can use dynamic analysis to mine the sensitive information in the inference codes as the correct model information and intermediate results must be recovered at runtime for static and half-dynamic obfuscation methods. We assess the vulnerability of the existing obfuscation strategies using an instrumentation method and tool, DLModelExplorer, that dynamically extracts correct sensitive model information at runtime. Experiments show it achieves very high attack performance. To defend against such attacks based on dynamic instrumentation, we propose DynaMO, a Dynamic Model Obfuscation strategy similar to Homomorphic Encryption. The obfuscation and recovery process can be done through simple linear transformation for the weights of randomly coupled eligible operators, which is a fully dynamic obfuscation strategy. Experiments show that our proposed strategy can dramatically improve model security compared with the existing obfuscation strategies, with only negligible overheads for on-device models.

A high-level modeling language for the efficient design, implementation, and testing of Android applications

Designing a 3D modelling tool for novice users.

ANDROID CODE AUTOMATIC GENERATION BASED ON APPLICATION DESCRIPTION

Adroitness: An Android-based Middleware for Fast Development of High-performance Apps

DEVSMO: an ontology of DEVS model representation for model reuse

Development Framework for Mobile Application Based on MDA

A Mda Based Approach For Multi-Platform Mobile Application Development

A Modeling Tool for Sensor-based Mobile Applications

Assistance in Model Driven Development: Toward an Automated Transformation Design Process

Command Pattern-based MOF Design and Testing

[email protected]: representing run-time system data as MOF-compliant models

DynaMO: Protecting Mobile DL Models through Coupling Obfuscated DL Operators

Refactoring Android Java Code for On-Demand Computation Offloading

MOF compliant fundamentals for multi-domain system modeling and simulation

HealMA: a model-driven framework for automatic generation of IoT-based Android health monitoring applications

An Approach for Modeling and Analyzing Code Mobility.

Umple: Model-driven development for open source and education

Applying MDE Tools at Runtime: Experiments Upon Runtime Models

DroidMD: an Efficient and Scalable Android Malware Detection Approach at Source Code Level

From Executable UML Model to J2ME Program

ANDROID APPLICATION ASSEMBLY TECHNOLOGY STUDY BASED ON CODE SNIPPET REUSE