(Expired) May 2022 - Master's Thesis - Learning how to drive in CARLA

Topic: Learning how to drive in CARLA

The Robot Learning and Deep Learning groups (led by Prof. Joni Pajarinen and Prof. Alexander Ilin) are seeking a talented student to write a Master’s thesis in the area of machine learning for autonomous driving. In this project, we will learn how to drive in the CARLA simulator (https://carla.org/), we will explore the limitations of the current state-of-the-art (see, e.g., [1]) and develop new learning algorithms. The developed methods may be based on object-centric unsupervised learning [2], reinforcement learning, planning under uncertainty, or multi-agent reinforcement learning [3]. The ultimate goal is to learn driving policies which are capable of handling events that happen rarely in real life.

[1] Chen and Kähenbühl. Learning from all vehicles, CVPR 2022.
[2] Kipf et al. Conditional Object-Centric Learning from Video, ICLR 2022.
[3] Strouse et al. Collaborating with Humans without Human Data, NeurIPS 2021.

🎓 Supervisors: Joni Pajarinen, Alexander Ilin
📍 Aalto University, Finland
⚡️ Salary: 2335 EUR/month
⏳ Deadline: 14 June 2022