Lukas Herron

This paper introduces REDIAL, an unsupervised framework to detect overparametrization in RNA language models, revealing inefficiencies and design flaws in current models.

DiffDock-Glide improves molecular docking by combining a diffusion model with Glide's post-docking minimization, showing better performance on novel targets.

RNAnneal predicts RNA structural ensembles using deep learning and physics-based modeling, outperforming existing methods in classifying experimentally resolved conformations.

This paper introduces a generative model based on the consumer/resource framework to design host-associated microbiomes that correspond to desired host phenotypes and vice versa.

SMbiot identifies low-dimensional latent variables that capture host-microbiome associations and predicts missing metadata or community compositions.