Simone Bianco

The paper introduces a deep learning framework to resolve mechanistic ambiguity in complex signaling networks by learning dynamic signatures from competing models.

The paper shows that the choice of projection basis affects the maximum achievable performance in perturbation prediction, with different bases being optimal for chemical vs genetic perturbations.

This paper introduces a physics-based model explaining how cells maintain histone methylation patterns after division, validated by simulations and experimental data.

The paper uses synthetic costimulatory domains and machine learning to identify design rules for CARs that enhance cytotoxicity and stemness.

This study develops a deep learning model to predict phenotype probability distributions from noisy biological data with few observations per condition.

The paper predicts epitope candidates for SARS-CoV-2 by analyzing epitope sequences from related coronaviruses and identifying conserved and diverse antigen regions across 62k genomes.

This paper develops a framework combining CNN and unsupervised learning to map genotype to vacuole morphology in yeast, identifying 4 phenotypic groups with 85% purity and 80% prediction accuracy.

The paper develops a semi-supervised pipeline to annotate SARS-CoV-2 genomes, identifying molecular targets with high accuracy and scalability.

This paper applies active deep learning to reduce annotation effort in cellular segmentation, showing reduced training samples and improved segmentation quality.

The paper investigates the role of nitric oxide signaling in cardiac progenitor differentiation and heart tube formation in chicken embryos using computer vision for cell counting.

This paper introduces algorithms for automatic plankton classification and anomaly detection with minimal supervision, achieving performance comparable to supervised methods on real-world datasets.

Quantius provides a crowd-based image annotation platform that achieves high-fidelity human annotation at 105 clicks per hour, matching deep learning performance while reducing time and cost.