Formal Verification of Deep Neural Networks for Code Conversion applications | FVDNNCC
Cooperating countries: Egypt and Austria
Coordinating institution: Vienna University of Technology
Project coordinator: Nahla El-Araby
Partner institution: School of information technology and computer science of Nile University
Project duration: 01.02.2024 - 31.01.2027
Project summary
Mobile Applications are very crucial nowadays in our lives. Many applications have been developed and used worldwide. Many of those applications can be important to human life in different ways like disabled aids, elderly people support, smart home controls, educational platforms and remote resource access controls. Since many applications need to run on various platforms, there exists a great need for cross platform mobile applications that can easily execute on different devices. Mobile applications can be translated from one code language to another through code conversion tools. A smart accurate code conversion framework would ease the process of developing applications for different operating systems and help smooth migration to updated hardware technologies with better resources. Automatic code translation would save designers efforts and lead to power efficient and memory efficient translated programs.
In this project, we aim to develop a tool for language conversion that can be used to easily map mobile applications into different platform. From our previous research, it has been proved that this tool can be more effective by incorporating Deep Neural Networks (DNNs). Also, formal verification can greatly enhance the quality of code conversion tools and decrease the design time. Incorporating formal verification with DNNs would lead to an acceptable level of trust in the neural network decisions. The aim of this work is to provide a smart verified correct by construction code conversion platform.