AI Paper Spotlights - Self-Supervised Transformers for Drug Discovery

Join us for the upcoming round of the AI Paper Spotlights - Berlin’s AI Research Meet Up as we dive into the latest advancements in the dynamic world of AI 🚀

‍

Self-supervised learning (SSL) leverages vast unlabeled datasets to uncover powerful representations that can be harnessed for a broad spectrum of generative and discriminative tasks. In this talk, we’ll explore how self-supervision, a key driver of large language models and recent advancements in computer vision, can improve image-based compound screening and advance early drug discovery. In particular, we’ll delve into self-supervised compound profiling at whole-image and single-cell levels and briefly review the SOTA methods that combine self-distillation and masked image modeling.

‍

Join us on the Merantix AI Campus, where you can connect with the thriving Berlin ML community, or alternatively, join us virtually!