If you are interested in any of these projects, please email your CV and a brief transcript to Dr. Behzad Bozorgtabar at behzad@ece.au.dk.
1. Efficient Test-Time Adaptation for Edge Intelligence
Objective: Develop low-latency and memory-efficient test-time adaptation for edge perception systems such as drones, under changing weather, lighting, altitude, and scenery.
Requirements: Strong Python/PyTorch, deep learning for CNNs/ViTs, and interest in TTA, model efficiency, or embedded AI.
2. Parameter-Efficient Test-Time Adaptation for Tiny Vision-Language-Action Models
Objective: Develop lightweight test-time adaptation for tiny vision-language-action (VLA) models to improve robotic manipulation under deployment shift.
Requirements: Strong Python/PyTorch, transformers/VLM knowledge, and interest in LoRA, TTA, or robotics/embodied AI.
