Problem & Setting

Using prior knowledge, Few-Shot Learning aims to (rapidly) generalize to new tasks containing only a few samples of information.

We have multiple possible projects available around this topic, which we're happy to discuss once you reach out (arnout.devos@epfl.ch).

Research work from our group in relation to few-shot learning:
Model-Agnostic Learning to Meta-Learn
A meta-learning approach for genomic survival analysis
Self-Supervised Prototypical Transfer Learning for Few-Shot Classification
Regression Networks for Meta-Learning Few-Shot Classification
Reproducing Meta-learning with differentiable closed-form solvers

Required skills

• Interest in machine learning research
• Strong programming skills (Python & PyTorch preferred)

Benefits

• Explore the active field of research in few-shot learning
• Get hands-on experience with state-of-the-art machine learning methods and libraries