Bernhard Schölkopf

This paper introduces RigidSSL, a geometric pretraining framework for protein design and conformational modeling, improving designability and capturing realistic dynamics.

This paper introduces Decode-gLM, a toolkit using sparse autoencoders to interpret, audit, and steer genomic language models, revealing biological features and identifying data contamination.

The paper introduces eDICE, a deep learning model for imputing epigenomic tracks, showing improved performance on reference data and successful application to individual-specific epigenomic patterns.

ResMiCo improves metagenome assembly quality by accurately identifying misassembled contigs using deep learning, achieving better performance than existing methods.

The paper evaluates 23 imputation models for cross-cell type epigenomic data, finding quality assessment challenging due to distributional shifts, data availability, and redundant metrics.

DeepMAsED evaluates metagenomic assembly quality using deep learning, identifying misassembled contigs without references and showing high accuracy on large assemblies.

GeNet provides deep representations for metagenomics classification, achieving competitive precision and recall with lower memory use and enabling effective downstream tasks.