TUM Open Day

The MCML at Open Day Campus Garching

Experience research up close for a whole day

   03.10.2024

   10:00 am - 5:00 pm

   TUM, Boltzmannstraße 3, 85748 Garching bei München

Can AI learn to see? Gain insights into the research and test it yourself! With the AMELIE exhibit, the MCML invites visitors to learn about the basics of machine learning in a playful and interactive way.

In a second demo, we show how it is possible to simultaneously localize and map the environment using a camera and a connected IMU sensor. The method is called DM-VIO and can be used in self-driving cars, for example.

We also provide an insight into Stable Diffusion, a deep learning model that generates images from text descriptions. We explain how Stable Diffusion learns the connection between image and text through a latent diffusion model process and invite you to try out Stable Diffusion for yourself.

On a big, illustrated canvas we show how AI works and explain the difference between AI and Machine Learning. Discover the basics of this technology and find out how it is impacting our everyday lives.

Location: Boltzmannstr. 3

---

Our PI Felix Krahmer will be giving a lecture titled “Why do simple training methods in artificial intelligence work so well?” (Lecture in German)

Short Info:
In machine learning, complex neural networks are often used and trained with large datasets. However, the training methods are usually relatively simple. This lecture aims to provide insight into how mathematics can help explain why these methods still perform so well.

Location: Lecture Hall 1 of the Mathematics/Computer Science Building, Boltzmannstr. 3.

---

Two scientists from the MCML will be presenting their research during the science slam events of the Open Day at 12:00 and 14:00.

Location: MW1801, Maschinenwesen, Boltzmannstr. 15

---

Amelie was built by the LMU media informatics group of MCML PI Albrecht Schmidt. DM-VIO is a project of the TUM Computer Vision Group led by MCML Director Daniel Cremers. Stable Diffusion was developed by the LMU Computer Vision & Learning Group of MCML PI Björn Ommer.

---