Automated Software Engineering Seminar
- ÕIS data: MTAT.03.271 (PLRS), 3 ECTS.
- Time: Tuesdays, 16:15 - 18:00
- Place: Liivi 2-612
- Language English
- Contact: Vesal Vojdani
The primary goal of this seminar is to read and understand key papers in the field of automated software engineering. We focus specifically on the emerging field of Deep Program Understanding that aims to teach machines to understand complex algorithms, combining methods from the programming languages and the machine learning communities. The seminar will consist of background lectures and student presentations.
Requirements
To pass this course, you are expected to study a research paper and give a self-contained 90 minute lecture about it. The paper should be in the field of automated software engineering, specifically you can choose topics from the following two sources, or pick any paper that is similar in spirit to those mentioned here:
- Deep Program Understanding. This is a Microsoft Research project that focuses on using machine learning to learn algorithms from examples, automate software engineering tasks, and improve program analysis. All their papers are excellent fits for this seminar.
- DiffBlue Research. DiffBlue is a self-described world leader in "AI that understands code". They use techniques from computer-aided verification and machine learning. Again you are welcome to select any topic from their research page.
In addition to your presentation, you need to analyze two of the paper that are presented by other students. For your chosen papers, you should write a brief summary and invent three relevant questions that you will ask the presenter.
Implementation Project
If you want to try the ideas in the paper, or maybe you just want do some practical work during this seminar. Well, we can register you to MTAT.03.272 (Special Assignment in Programming Languages). This provides an extra 3 ECTS credits for completing an independent task in the field of automated software engineering.