Events

kennedy+swan: The Red Queen Effect

Starting January 26, 2026, the foyer of the Rahel Hirsch Center will host a new exhibition dedicated to a topic central to research at BIH and Charité: artificial intelligence.

BLISS Speaker Series #30 - "The Price of Intelligence: Three Risks Inherent in LLMs"

In the BLISS AI Speaker Series Ahmed Salem who is currently a Researcher at Microsoft Security Response Center (MSRC) and will discuss "The Price of Intelligence: Three Risks Inherent in LLMs", lasting approximately 45 minutes. After the talk, seize the opportunity to connect with fellow AI enthusiasts to share ideas and questions while enjoying free drinks and pizza.  

HIDA Lecture: When AI Performance Misleads

Despite impressive performance in publications, many AI models fail when deployed in real-world settings. One important reason is poor or misleading validation.

Deep Learning Models of Regulatory DNA: A Comparison of Model Design Choices

Chromatin state and gene expression is tightly regulated by proteins that interpret sequence syntax encoded in regulatory DNA. Genetic variants influencing traits and diseases often disrupt this syntax. Several deep learning models have been developed to decipher regulatory DNA and identify functional variants. Most models use supervised learning to map sequences to cell-specific regulatory activity measured by genome-wide molecular profiling experiments.

"The Target-Disease Model Is Dead. AI Can’t Fix It — Yet"

The target-disease model that built modern pharma is dead — not because AI killed it, but because biology walked away from it. As molecular mechanisms are mapped with increasing granularity, biological reality reveals itself as combinatorial, redundant, and fragmented. 

Lunch Talk: Leonard Hackel "Towards Efficient Foundation Models in Remote Sensing"

Leonard Hackel from the Big Data for Earth Observation and Remote Sensing Image Analysis group will talk about "Towards Efficient Foundation Models in Remote Sensing". In this talk Leonard Hackel will propose a soft mixture-of-experts masked autoencoder for remote sensing foundation models that more than doubles computational efficiency while maintaining competitive performance across downstream tasks and enabling scalable, cross-modal learning.

Cedalion Workshop

This scientific workshop introduces researchers to Cedalion, a Python-based toolbox for advanced fNIRS and DOT analysis in naturalistic settings. Built on HOMER3 and AtlasViewer, it integrates machine learning and multimodal data processing for cutting-edge neuroimaging research.