Promotion - Effiziente Neuronale Repräsentation von Datensätzen
Core
Researching deep generative models to create efficient, on-demand data representations for training and validating downstream neural networks.
Role type
PhD researcher (Deep Generative Models)
Builds
Novel learning algorithms for synthesizing relevant data on demand to improve training efficiency and ensure desired invariance.
Domain
Artificial Intelligence, Deep Learning, Computer Vision
Deliverable
research
Required skills
Deep Learning, Computer Vision, Python, Deep Generative Modeling, Foundation Models
Preferred skills
Peer-reviewed publications, experience with Diffusion Models, GANs, VAEs
Technologies
TensorFlow, PyTorch
Responsibilities
Develop new approaches to adapt deep generative models as data sources; utilize controllability of generative base models to treat datasets as structured representations; discuss and develop new ideas with Deep Learning and Computer Vision experts; publish results in high-ranking journals and conferences.
