Till Richter

This paper develops a continual learning framework to map colorectal cancer cell states to causal mechanisms through integration of single-cell data and perturbation experiments.

This paper introduces a metric to evaluate synergistic integration in multimodal cell models and shows when it benefits tasks over unimodal approaches.

CellHermes integrates multimodal omics data using a biological language model, achieving comparable or better performance than single-cell models across various tasks.

This study benchmarks self-supervised learning techniques in single-cell genomics, finding masked autoencoders outperform contrastive methods in cell type prediction and data integration.