Software plays a crucial role in the society, influencing various aspects of daily lives. During the development of a software, writing source codes is the core activity for software engineers. However, developing reliable codes is costly because it requires constantly references to documentations and online resources and making sense of the logic behind existing code bases. These are both challenging and could slow down the developing process.
We are motivated by this and would seek to model the relations between natural language queries and source code snippets in order to help increasing the efficiency of software development.
This capstone project is interdisciplinary, that requires us to apply skills or investigate issues across many different Ischool subject areas such as Machine Learning, Natural Language Processing and Web Application. In addition, we hope Code Search to be a user-friendly application, so we will consider not only the model accuracy but the user experience to meet the needs of all kinds of programming users.