Thema der Dissertation:
MACHINE LEARNING METHODS FOR DNA METHYLATION-BASED CLASSIFICATION OF TUMORS Thema der Disputation:
Single-Cell Foundation Models as Models of Biology
MACHINE LEARNING METHODS FOR DNA METHYLATION-BASED CLASSIFICATION OF TUMORS Thema der Disputation:
Single-Cell Foundation Models as Models of Biology
Abstract: What does it mean for a foundation model to be a model of biology? Single-cell foundation models learn representations from vast collections of transcriptomes, yet how much biological understanding these representations encode remains unclear. This talk considers where the limitations lie — in model architecture, in evaluation, and in the pretraining signal itself — and asks whether the integration of multiple modalities offers a way forward.
Zeit & Ort
14.08.2026 | 14:00
Seminarraum 007/008
(Fachbereich Mathematik und Informatik, Arnimallee 6, 14195 Berlin)