Romain Menegaux

This study evaluates foundation models for predicting cellular responses to perturbations, finding that some pretrained embeddings improve predictions while others do not.

Phylo-HS improves taxonomic classification speed and accuracy by using a hierarchical softmax method that leverages phylogenetic structure.

The paper introduces Brume, a method that reduces memory usage for training compositional machine learning models with longer k-mers in metagenomics, achieving better performance on benchmarks.

The paper introduces fastDNA, a model that embeds DNA sequences into a vector space for metagenomics classification, achieving higher accuracy and scalability than existing methods.