Korea University

Single-cell foundation model pretrained on human brain organoids that predicts transcriptome-wide responses to knockdown of any protein-coding gene.

EEG foundation model that decodes by matching neural activity to label text embeddings, with one instruction-tuned checkpoint covering seven tasks.

Gastric pathology model reading whole-slide images through a chain of dependent questions, mirroring a pathologist's stepwise reasoning.

Open-weights pathology foundation models pairing a 1.1B-parameter ViT-g/8 tile encoder with distilled 86M and 22M variants for H&E histology.

RNA small-molecule binding site prediction from sequence alone, pairing frozen RiNALMo embeddings with a lightweight MLP classifier.

Hierarchical immune cell type annotation for scRNA-seq, rendering expression as images for a CNN. 93.2% mean subtype accuracy over seven datasets.

Pathology video-language model that reads histopathology clips and produces a step-by-step histological description plus a sign-out diagnosis.

Self-supervised pathology foundation model with a 300M-parameter vision transformer tile encoder, trained on a multi-stain whole-slide image corpus.