Benjamin Fefferman

bfefferman [at] g.harvard.edu

I work on foundational AI problems — computational, statistical, and algorithmic — with biomedical and societal relevance. I am particularly interested in research topics pertaining to precision oncology and drug discovery, for which I pursued Ph.D. studies in bioinformatics.

My research has since evolved toward interdisciplinary questions pertaining to how we can ensure that the AI revolution benefits humanity. I recently received my Master’s degree from Harvard University, and have since been awarded a fellowship to conduct AI safety research in Cambridge, UK.

I received my Bachelor of Science from the University of Chicago, where my research centered about chemical biology for epigenomics & epitranscriptomics, as well as machine learning.

My prior studies and current work involve extensive context from the mathematical, natural, and data-driven sciences.

selected publications

DiffeoMorph: Learning to Morph 3D Shapes Using Differentiable Agent-Based Simulations

arXiv preprint arXiv:2512.17129, 2025

Abs HTML

We introduce DiffeoMorph, an end-to-end differentiable framework for learning a morphogenesis protocol that guides a population of agents to morph into a target 3D shape. Each agent updates its position and internal state using an attention-based SE(3)-equivariant graph neural network, based on its own internal state and signals received from other agents.
LABS: linear amplification-based bisulfite sequencing for ultrasensitive cancer detection from cell-free DNA

Genome Biology, 2024

Abs DOI HTML

We report LABS, which enables linear amplification of bisulfite-treated DNA fragments in a genome-wide, unbiased fashion, detecting cancer abnormalities with sub-nanogram inputs. Applying LABS to 100 patient samples revealed cancer-specific patterns, copy number alterations, and enhanced cancer detection accuracy.