Data Augmentation: From Hand-Crafting to Generative Models

In this talk, we will take a close look at strategies for data augmentation in machine learning. While data augmentation started out as a set of simple transformations in Computer Vision, it has grown into its own field of investigation, including fields such as NLP and Robotics. Modern approaches rely on diverse augmentation strategies and even learned generative models to produce high-quality augmentations. We will discuss major milestones of this development, as well as pros and cons of different approaches.

About Dr. Yannik Kilcher

Yannic has a BSc in CS from the University of Basel and a MSc and PhD in Machine Learning from ETH Zurich. He is a co-founder at DeepJudge, a NLP startup for LegalTech, and also maintains a popular YouTube channel on various topics in AI research.

Useful Links

Yannik YouTube Channel: https://www.youtube.com/channel/UCZHmQk67mSJgfCCTn7xBfew

Registration

The seminar will be held online via Zoom on June 18th at 14.30 CET.

To attend the seminar please send an email to [email protected]