Human compatible world models across sizes, languages and modalities

Did you know there was an AI startup working on "General Artificial Intelligence" right here in Heidelberg?
Yes, there is! And we will have them with us at heidelberg.ai!
We are excited to have Jonas Andrulis, Constantin Eichenberg and Robert Baldoch, working at Aleph Alpha from Heidelberg, present their work and insights on multimodal learners and world models in person at the DKFZ.
As an example of their work and goal towards making AI technologies accessible to a wide audience, they open-sourced the MAGMA multimodal model, which can process images and text in any combination, this year.
For a full abstract head over to https://heidelberg.ai/2022/09/27/andrulis.html . Drinks and snacks will be provided, courtesy of the Division of Medical Image Computing and Interactive Machine Learning Group at DKFZ.