발간년도 : [2023]
| 논문정보 |
|
| 논문명(한글) |
[Vol.18, No.4] Construction of eGovernment Framework Code Generation Model Using BERT |
|
| 논문투고자 |
Jiho Song, Panyoung Kim, Inbin Choi, Hoekyung Jung |
|
| 논문내용 |
There are many difficulties in writing code while performing various projects based on the e-government framework. Therefore, artificial intelligence(AI)-based software development tools, various cases of converting natural language used when talking to people into a programming language that computers can understand are increasing. By using OpenAI's Codex model to learn the source code owned by Github, developers install GitHub Copilot Extension in an integrated development environment(IDE) and enter annotations or function names, automatically completing and presenting the source code to be written by artificial intelligence to increase developer productivity. However, for codes that artificial intelligence has not learned, it does not present perfect codes, and sometimes AI misunderstands the developer's intentions and presents codes that need to be modified in large part. If that happens, it could rather negatively affect productivity. In this paper, we propose a code generation model that automatically generates the source code by learning the source code used in projects based on the e-government standard framework to the BERT model, as Github Copilot learned the code using the Github repository. Through this, it is believed that it can be used as a tool to improve the productivity of the project, reducing code writing time, and accumulating know-how in the development process in the model to be used in various application cases as a communication tool for developers. |
|
| 첨부논문 |
|
|
|
|
|
자료실 > 논문지 
