Master Thesis, 30 HP: AI/ML for automatic program repair in C2 mission systems
We are looking for two students that will help us explore the latest possibilities in Artificial Intelligence and Machine Learning.
We build and maintain software systems for our Combat Systems & C4I Solutions. Build, test and packaging are fully automated and our continuous integration engine runs frequently. The software development effort engages several teams, with distributed responsibilities, collaborate to integrate their respective parts into complete software systems. Failures detected late in the process, e.g. in the integrated software system, are often hard to locate in terms of bugs. One reason, of course, is that there can be hundreds of commits since the error prone code was introduced. These incidents stalls the development effort as no new features are allowed to enter the system, are cost-wise in terms of resources needed to detect and fix the error, and, in worst case, can jepordize important mile stones towards customers.
Description of the master thesis
This Master Thesis aims at providing a sound and thorough assessment of the usability of state-of-the-art repair tools in the context of our Continuous Delivery Pipeline. We want to
- survey the existing automatic program repair tools
- integrate a state of the art tool into our CI
- establish a corpus of bugs and patches (e.g. from previous commits)
- determine to what extent the tool is able to reproduce the manual patches
- assess the scalability of existing tools
Please find an overview of a Continuous Delivery Pipeline here: https://www.scaledagileframework.com/continuous-delivery-pipeline/.
Relevant literature includes:
- How to Design a Program Repair Bot? Insights from the Repairnator Project (Simon Urli, Zhongxing Yu, Lionel Seinturier and Martin Monperrus), In 40th International Conference on Software Engineering, Track Software Engineering in Practice, 2018
- A Critical Review of "Automatic Patch Generation Learned from Human-Written Patches": Essay on the Problem Statement and the Evaluation of Automatic Software Repair (Martin Monperrus), In International Conference on Software Engineering, 2014
There is also an interesting topic on SVT play (starting at 0.52): https://www.svt.se/nyheter/inrikes/de-lyfter-sveriges-ai-satsning
This Master Thesis is suitable for 2 students with interest in DevOps, Continuous Delivery, Docker, Python, and of course, machine learning. You are at the end of your computer science studies, or equivalent, and about to start your Master Thesis work for 30 HP.
We hope to get acquainted with communicative, curious and positive students with great passion who want to help us explore the latest technologies and trends in the software field.
What you will be a part of
We are building Combat Systems & C4I solutions for every type of naval platform, ranging from combat boats and patrol boats, to frigates and aircraft carriers, as well as submarines. Continuous Delivery and DevOps are keywords for us when aiming for increasing product quality and shortening the release cycle. In that spirit we now want to evaluate techniques in AI/ML for automatic program repair.
Last application day
Thomas Lindén, Manager
073-437 52 39
If you aspire to help create and innovate whilst developing yourself in a challenging team setting, Saab may well have the perfect conditions for you to grow. We pride ourselves on a nurturing environment, where everyone is different yet we share the same goal – to help protect people.