From Model-centric to Data-centric AI

October 26th, 09:30 – 10:15 am

An AI model is only as good as the data that feeds it. So far, AI and machine learning projects have focused on the models. Primary goal was to improve the performance of a model in its respective use case by evaluating methods, tuning parameters etc. A new trend wants to shift the focus to the data. In his presentation Walid Mehanna will talk about why focusing on the data in the first place will pay off when heading towards scaling AI and MLOps. He will also elaborate on how data-centric AI can help to increase the accuracy and ethical standards of your AI endeavors. In the panel we …

… elaborate how focusing on the data in the first place will increase the value of your AI model.

…discuss in which areas and industries the data-centric approach is more prevalent than in others.

…debate how a systematic approach to data-centric AI might look like.

…give tips how to find the right balance between model-centric and data-centric AI.

In this panel, there will be a 10-minute introductory presentation by Walid Mehanna followed by a 30-minute discussion.

Moderator:

Dr. Carsten Bange, Founder and CEO | BARC GmbH

Speakers:

• Walid Mehanna, Group Data Officer | Merck Group
• Dr. Lydia Nemec, Head of AI Accelerator | Carl Zeiss AG
• Divya Bokaria, Head of Data | Zattoo