Beneficiary names: AnteP92 Informatics Consulting and Service Limited Liability Company, FrontEndART Software Ltd.
Project title: "Development of a Cloud-Based Measuring and Analysis Service Supporting Machine Learning-Based Improvement of Software Source Code Quality"
Contracted support amount: HUF 214,749,668
Amount of support awarded to FrontEndART Software Ltd.: HUF 77,864,941
Support rate (%): 56.55%
Project description:
A two-member consortium was formed to implement the project titled "Development of a Cloud-Based Measuring and Analysis Service Supporting Machine Learning-Based Improvement of Software Source Code Quality." The consortium aims to carry out the development detailed in the project with eligible costs of HUF 379.7 million, financed by a non-refundable grant of HUF 214.7 million. The consortium leader, AnteP92 Ltd., offers sophisticated software solutions in the market, focusing on the development of key business applications for the media, mobile phones, telecommunications, financial services, and automotive industries. The consortium partner, FrontEndART Ltd., was founded by researchers from the Department of Software Engineering at the University of Szeged who specialize in source code analysis. The company provides services in quality-managed software development, software maintainability education, software risk assessment, and monitoring.
The goal of this project is to develop a virtual assistant based on machine learning that facilitates the improvement of software maintainability. The assistant, named CODEE, will be developed and marketed as a cloud-based service. CODEE will continuously learn using a deep-learning algorithm on a training database based on Big Data technologies, specifically created and maintained for this purpose within the project. The structured data, which impacts the maintainability of thousands of software systems and millions of software versions, will be produced by automatic analysis tools relying on open sources. In the future, during the sale of the service, the training database will also be expanded with measurement data from closed-source systems. The so-called agent applications responsible for data collection will continuously monitor various information and data sources available on the internet. The relevant and accessible information will be downloaded, processed, analyzed, and uploaded to the central training database in an appropriate structured format.
Project start date: 2021.01.01
Project physical completion date: 2022.12.31
Project ID: 2020-1.1.2-PIACI-KFI-2020-00078