10th #MLDD - But WHY?!

Hi Everyone,

we are delighted to announce the 10th (!!) #MLDD: „But WHY?!“. We’ll meet on Wednesday, November 30th at the elevait HQ right on Albertplatz.

We are more than happy that we could win Peter Steinbach, Lead of the AI Consultant Team at Helmholtz AI for this special occasion. His team's mission "[...] is to consult scientists primarily of the research field Matter in the application of automated data processing and knowledge extraction methods. We want to disseminate state-of-art best practises in ML and data science. With this, we hope to boost data understanding of our clients at the global academic scale in order to provide a competitive advantage. Within this mandate, we will try to advance methods or tooling in order to reduce the time investment on our as well as on our clients' side."
His talk will focus on the explainability of models - a topic which might be one of the biggest concerns when it comes to a more widespread application of AI in many real-world applications.

As always, there will be time to connect after the event including a fine selection of snacks and drinks.

But that's not all. Following the discussions after our last event, we've created a small Slack community which will hopefully serve as a vibrant platform for exchanging ideas for this event in the future - independent of whether we meet in person or not.

Join

See you there :)
Cheers Stefan & Alex

-- EDIT (25/11/2022) --

Hi Everyone,

there are two important changes to our event.

1. There is a second talk \o/: Gregor Blichmann (CTO @elevait) und Ege Tuncer (Trainee Machine Learning @elevait) will present their findings on "How to train handwriting recognition without training data".
2. Peter will talk about a slightly different topic than what we announced initially - but do not be afraid - it only gets better :)

We are really excited about this event.

See you Wednesday!

Read the two abstracts
-------------------------------
"How to train handwriting recognition without training data?"
We have all heard that a handwritten letter is more personal than a text message ... but is it?
There is a whole forensic science branch concerned with finding out if a person wrote the text in question or not ... but will this science and craft stand the test of time?
AI is quickly and steadily transforming many of what we consider uniquely human. To some, handwriting is an individual, recognizable human trait. But we want to show you an open-source repository, which might make you answer the above questions differently and give you some food for thought on what that means and how you could use it.

Speaker:
Gregor Blichmann (CTO @elevait) und Ege Tuncer (Trainee Machine Learning @elevait)

----------------------------------------------------------------------------------------

In this presentation, I'd like to take the audience on a tour of Helmholtz AI at HZDR. This will encompass a brief recap of our mandate, our structure and spotlights of current work undertaken in both Helmholtz AI teams - consultants and scientists alike.
Further, I will use the latter half of the presentation to introduce the audience to (group) equivariant neural networks. This new field of machine learning aspires to combine prior knowledge about symmetries in training data and established machine learning training techniques. I will motivate this technique, show results from the community and finish with a discussion of its application to semantic segmentation in images.

Speaker:
Peter Steinbach (Lead AI Consultant Team @HZDR)