Zeyu Fu

CLOP-DiT generates synthetic single-cell gene expression profiles from structured biological metadata using contrastive alignment and diffusion models, achieving high steering accuracy but limited cross-dataset variability.

This paper introduces DPMM-Base and DPMM-FM autoencoders with Dirichlet Process priors for single-cell representation learning, showing improved geometry but reduced label recovery accuracy across multiple scRNA-seq datasets.

This paper introduces Topic-FM, a neural topic VAE for single-cell transcriptomics that improves interpretability and clustering performance across multiple datasets.

MoCoO improves single-cell trajectory inference by combining VAE, Neural ODE, MoCo, and Flow Matching, achieving top performance in clustering and embedding quality across multiple datasets.

This study introduces scRL, a reinforcement learning method for assessing cell fate decisions in single-cell data, showing superior performance compared to existing techniques.