Supervision: Ali Hariri

Project type: Semester project (master)

Available

Project Summary: In this work we address the scalability issues of point cloud methods. We will aim to develop more efficient Deep Learning algorithms for point cloud processing through Graph Neural Networks (GNNs). The goal is to explore novel methods of message passing on point clouds to make the learning process more scalable and less memory intensive. To proceed, we inspire from recent work on Graph Transformers.

Required Skills: The student should have a solid understanding of Deep Learning methods. Prior experience working with deep learning frameworks such as PyTorch or TensorFlow would be highly preferable, in addition to existing knowledge about network machine learning or Transformer architectures.

To apply for this project, please send me a copy of your CV and transcript of grades at ali.hariri@epfl.ch, happy to chat further if you're interested !