Zsolt Kira

DiffSeg offers a new way to segment images without labeled data.

Introducing a fresh approach to enhance machine learning model performance with limited data.

Semantic Placement enhances AI's ability to place objects based on context.

Examining the links between image generation models and associative memory.

A new benchmark to improve how robots locate objects using various methods.

Learn how selective training can improve robot learning efficiency and adaptability.

Learn how adversarial attacks manipulate deep learning through differentiable rendering techniques.

GRAIN improves image understanding by aligning detailed descriptions with images.