I’m a PhD student advised by Steve Zdancewic and Konrad Körding. My research interests are in programming language theory, neural networks, and cognitive science.

My work focuses on identifying programming languages which can be implemented by neural networks—building compilers that instantiate programs as trained neural networks and proving they are correct.

I also direct REPL, an NSF undergraduate research internship in programming language theory. It was designed to encourage historically marginalized groups to pursue PhDs in programming language theory.

---

Research

Theses

Compositional desires as compositional programs. MSc Thesis, MIT (2021).
Learning in System F. MSc Thesis, University of Oxford (2019).

Publications

Research experiences for undergraduates are necessary for an equitable research community
Joey Velez-Ginorio, Joshua Sunshine (equal contribution)
Communications of the ACM, Volume 67, Issue 8 (2024)

Effects and coeffects in call-by-push-value
Cassia Torczon, Emmanuel Acevedo, Shubh Agrawal, Joey Velez-Ginorio, Stephanie Weirich
Proceedings of the ACM on Programming Languages, OOPSLA (2024)

When naïve pedagogy breaks down
Rosie Aboody, Joey Velez-Ginorio, Laurie Santos, Julian Jara-Ettinger
Cognitive Science (2023)

Towards neural functional program evaluation
Torsten Scholak, Jonathan Pilault, Joey Velez-Ginorio
AIPLANS Workshop, 35th Annual Conference on Neural Information Processing Systems (2021)

When teaching breaks down
Rosie Aboody, Joey Velez-Ginorio, Laurie Santos, Julian Jara-Ettinger
Proceedings of the 40th Annual Conference of the Cognitive Science Society (2018)

Interpreting actions by attributing compositional desires
Joey Velez-Ginorio, Max Siegel, Joshua B. Tenenbaum, Julian Jara-Ettinger
Proceedings of the 39th Annual Meeting of the Cognitive Science Society (2017)

Temporal order-based first-take-all hashing for fast ADHD detection
Hao Hu, Joey Velez-Ginorio, Guo-Jun Qi
Proceedings of the 22nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining (2016)

Talks

When does X implement Y? Assessing Representation, Santa Fe Institute (2024).
Compiling to transformers. NJPLS, New York University (2024).
When do neurons* represent True? Kording lab meeting, University of Pennsylvania (2023).
When do neurons* represent True? Metauni festival, Online (2023).
Compilers as linking hypotheses. PLClub meeting, University of Pennsylvania (2022).
Compilers as linking hypotheses. Kording lab meeting, University of Pennsylvania (2022).
A neural compiler. Midlands Graduate School in Computing, Online (2021).
Finding programs in the brain. Fiete Lab meeting, Massachusetts Institute of Technology (2020). Compiler Synthesis. MFoCS Seminar, University of Oxford (2019).
Interpreting actions by attributing compositional desires. Cog & Dev. meeting, Yale (2018).
Interpreting actions by attributing compositional desires. CogSci, London UK (2017).