I'm an Assistant Professor of Physics at the University of Wisconsin-Madison, where I lead a group of researchers designing new AI methods to enable discoveries in particle physics and astrophysics.
I'm also a performing artist engaged in contemporary dance and theater. I lead research teams generating choreography with AI models trained on my own movements and integrate these models within my artistic practice.
Selected Scientific Projects
See my Google Scholar page for a complete list.
Multimodal Datasets with Controllable Mutual Information
A framework using flow-based generative models to create realistic multimodal datasets with precisely known mutual information, enabling systematic benchmarking of mutual information estimators and representation learning methods.
A Practical Guide to Unbinned Unfolding
A guide that collects practical advice from researchers across 11 major particle physics experiments that use OmniFold, a machine-learning-based unfolding technique for removing detector distortions from experimental data.
What's In Your Field? Mapping Scientific Research with Knowledge Graphs and LLMs
Using a large language model to extract structured knowledge representations and construct a knowledge graph of 30,000 arXiv papers across astrophysics, fluid dynamics, and evolutionary biology.
A 24-dimensional & unbinned measurement of Z+jets events
This precision measurement of high-momentum Z boson events uses neural networks to reduce detector distortions and facilitate direct comparison with theoretical QCD predictions.
Evidence of VH, H →ττ
The ATLAS Experiment's first evidence of this rare process of the Higgs boson decaying into two tau leptons.
Finding Stellar Streams
Using AI to identify remnants of ancient galaxies that can help map dark matter in the Milky Way, all without explicit labels.
Learning Likelihood Ratios with Neural Network Classifiers
An empirical evaluation of how neural network classifiers approximate likelihood ratios, tested on both multidimensional Gaussian distributions and simulated high-energy physics data.
Point Cloud AI Methods for Pion Reconstruction
Using Transformer, Deep Sets, and Graph Neural Network architectures to process pion calorimeter clusters and particle tracks as point clouds.
Symmetry Group Equivariant Architectures for Physics
We argue that the introduction of symmetries into an AI model's fundamental structural design can yield models that are more economical, interpretable, and/or trainable.
Creative Projects
Invisible Strings: Revealing Latent Dancer-to-Dancer Interactions with Graph Neural Networks
Using Graph Neural Networks to analyze contemporary dance duets, revealing partnership dynamics connecting the two dancers.
Dyads: Artist-Centric, AI-Generated Dance Duets
An attention-based Variational Autoencoder that generates a choreographic partner conditioned on an input dance sequence, developed through co-creation with dance artists.
mememormee
An AI-generated dance experiment and artistic residency at Amherst College.
- Spotlight presentation at the NeurIPS 2023 Workshop on ML for Creativity & Design.
PirouNet
Creating dance through artist-centric deep learning with a semi-supervised conditional recurrent variational autoencoder.
Kinetech Arts Shows
I was a dancer in two performances by Kinetech Arts inspired by entropy, AI, and technology.
- Sublimation, at David Ruth Glass Studio, Oakland, CA.
- Detour, at Zellerbach Playhouse, Berkeley, CA.
Cognicast Interview
An interview on the Cognicast podcast, hosted by technologist and musician Robert Randolph, about my research across AI, physics, and the performing arts.
Dancing With Myself
A talk at the StrangeLoop conference in St. Louis, MO, on AI, dance, and the creative process.
Untitled AI Birdsong Project
A 1-hour pop-up exhibit featuring AI-generated bird calls situated in nature.
Choreo-Graph
As an intern with Intel's AI Lab, I developed a Graph Neural Network to learn a latent graph representation of my dancing body.
Mirror Exercise
An AI-generated duet with myself.
- Featured at the 2020 NeurIPS AI Art Gallery, the AI Governance Forum, and the Boston Cyberarts Gallery.
SIGMA
A short film of entirely AI-generated movements.
- Featured at the 2019 NeurIPS AI Art Gallery for the Workshop on ML for Creativity and Design.
Beyond Imitation
I led a research project using variational autoencoders to generate choreography.