Natural language processing (NLP) has made considerable progress in the past few years, especially with transformer-based Machine Learning models. These advances have greatly powered improvements across various real-world applications, such as question answering, summarization, code generation (e.g., AlphaCode), chatbots (e.g., Amazon Alexa), drug discovery, or image generation from natural language (NL) description, and vice versa.
In this seminar, we will explore the latest research in NLP for source code, also known as Code Intelligence (CI), with a specific focus on transformers and structure injection.
Code Intelligence is an emerging field that applies NLP and Machine Learning to automatically analyze the source code of software and enables intelligent assistance to programmers.
We will review existing methods and benchmarks, and investigate downstream applications such as code generation, clone detection, code translation, code documentation generation, etc.
- Dozent*in: Thomas Arnold
- Dozent*in: Mert Tiftikci