SAM WITTY

I am the founder and principal consultant at Sorbus AI, where I help hard tech companies and R&D teams build and deploy AI models for their scientific and engineering challenges. Specifically, I help teams use state-of-the-art methods in causal inference and probabilistic programming that combine their existing mechanistic knowledge with data-driven machine learning.

Previously, I was a senior research scientist at Basis, where I was a core developer on the ChiRho causal probabilistic programming language. Before that, I completed my PhD at UMass Amherst and was a visiting researcher at MIT. When I'm not reading, writing, or talking about models, you'll likely find me lost in the woods with my family.

Contact me at sam[at]sorbus.ai

Selected Publications

For a complete list of publications, see my Google scholar.

Bayesian Structural Causal Inference with Probabilistic Programming
Sam Witty (2023).
PhD Thesis, University of Massachusetts Amherst [bibtex]

SBI: A Simulation-Based Test of Identifiability for Bayesian Causal Inference
Sam Witty, David Jensen, Vikash Mansinghka (2021).
arXiv preprint arXiv:2102.11761 [bibtex].

Causal Probabilistic Programming without Tears
Eli Bingham*, James Koppel*, Alexander Lew*, Robert Ness*, Zenna Tavares*, Sam Witty*, Jeremy Zucker* (2021, Alphabetical Order).
Third Conference on Probabilistic Progamming. [bibtex]

Causal Inference using Gaussian Processes with Structured Latent Confounders
Sam Witty, Kenta Takatsu, David Jensen, Vikash Mansinghka (2020).
International Conference on Machine Learning. [bibtex]

Bayesian Causal Inference via Probabilistic Program Synthesis
Sam Witty*, Alexander Lew*, David Jensen, Vikash Mansinghka (2020).
Second Conference on Probabilistic Programming [bibtex]