🔣 Input: your positive or negative sentiment.
📊 Output: A 2D visualization that shows how a Transformer model processes and transforms your words layer by layer.

See more 🔗 dcn.nat.fau.eu/ndw25-sentim...

#FAU #movingKnowledge #NdW25 #FAUMoD #DTNModConFlex #EUhorizon

This app makes those hidden transformations visible: as the layers advance, you can watch how your words become points on a plane, clustering and shifting as the model “understands” your sentence:
. . .

👉🏻Why it’s Interesting
Transformer models understand text by gradually updating internal word representations through multiple layers. Each layer refines meaning and relationships, but that process is normally invisible to us.

. . .

This post includes an app SentimentAnalysisTransformersApp created for a public outreach activity organized by our Chair during the Lange Nacht der Wissenschaften 2025 (Long Night Sciences 2025).

. . .

The authors of this post are supported by the @modconflex.bsky.social funded by the MSCA, European Union (Grant ID 10107355)

See more 🔗 dcn.nat.fau.eu/ndw25-sentim...

"How do #AI models understand emotions in text?

In this interactive experiment, you can type your own sentences and see how a transformer-based AI analyzes sentiment in real time. (...)"

SEE MORE 🔗 go.fau.de/1cbb-

#FAUDCNAvH #FAUMoD #longNightSciences #Sciences #AImodels #transformer

The workshop aims to advance research at the intersection of control theory, scientific computing, partial differential equations (PDEs), and machine learning.

SEE MORE 🔗 go.fau.de/1c-ec

We examine the intrinsic connections between representation, optimization, and control theory—rooted in cybernetics (from Ampère to Wiener) and historically motivated by the quest to design intelligent machines. (...)"

(...) This course explores these questions through an interdisciplinary lens, bridging PDE theory, control, and ML.

#FAU #movingKnowledge

About this course
▪️Can ML enhance PDE techniques without sacrificing mathematical rigor?
▪️ Can we develop hybrid computational frameworks that leverage data-driven approaches while maintaining the reliability of traditional methods?

in terms of moment order, terminal time, and measurement errors. In addition, we develop efficient numerical discretization schemes and demonstrate significant stability improvements of our approach through comprehensive numerical experiments."

🔗 cris.fau.de/publications...

#FAU #FAUMoD #research

transforming the problem into a more stable inverse moment formulation. Compared to existing methods, our moment-based approach reduces exponential error growth to polynomial growth with respect to the terminal time. We provide explicit error estimates on the recovered initial distributions
....

"We address the initial source identification problem for the heat equation, a notably ill-posed inverse problem characterized by exponential instability. Departing from classical Tikhonov regularization, we propose a novel approach based on moment analysis of the heat flow, (...)