Elizabeth Polgreen

Associate Professor
School of Informatics
University of Edinburgh

elizabeth.polgreen[at]ed.ac.uk

About

I am a Reader (~Associate Professor in US terminology) in the School of Informatics at the University of Edinburgh. I am interested in (formal and machine learning-based) synthesis techniques and the use of synthesis to increase the scalability of verification. I hold a research fellowship from the Royal Academy of Engineering.

Previously, I was a postdoctoral research scholar in Professor Sanjit Seshia's group at the University of California, Berkeley. I completed my PhD at the University of Oxford, supervised by Professor Alessandro Abate.

My CV is available here. A short bio for talk introductions is available here.

News

July 2026
  • I will be giving a keynote talk at SEFM 2026 in November.
June 2026
  • Zoey Shepherd will be talking at the SYNT workshop about Normalised Synthesis of Computations and our tool NSynC.
  • I was delighted to give the keynote talk at the Manchester conference for the Institute for Research in Schools.
May 2026
  • Our paper "Massively Parallel Mining of Specifications for Hardware Designs", led by Leiqi Ye and Guy Frankel, is accepted to appear at CAV 2026.
January 2026 August 2025
  • I have been promoted to Reader (~Associate Professor in US terminology).
  • We have two papers accepted to appear at FMCAD 2025: "PolyVer: a Compositional Approach for PolyGlot System Modelling and Verification" (preprint here ), led by Pei-Wei Chen and in collaboration with Professor Sanjit Seshia; and "Unlocking Hardware Verification with Oracle Guided Synthesis", led by Leiqi Ye.
  • Our paper "Guess, Measure & Edit: Using Lowering to Lift Tensor Code" is accepted to appear at PACT 2025
July 2025
  • I am part of a team of researchers that have been awarded a $5 million grant from DARPA as part of the TRACTOR program, Translating All C To Rust. This is in collaboration with Somesh Jha and Tej Chajed at The University of Wisconsin-Madison, Sanjit Seshia and Alvin Cheung at UC Berkeley and Varun Chandrasekaran at the University of Illinois Urbana-Champaign.
May 2025
  • Our paper, "HyGenar: An LLM-Driven Hybrid Genetic Algorithm for Few-Shot Grammar Generation" is accepted to appear in ACL findings 2025. Preprint here.
April 2025
  • Our paper, "Guided Tensor Lifting", is accepted at PLDI 2025. Preprint here. Yixuan Li will present this at PLDI in June 2025.
March 2025
  • Our paper, "Metric-Guided Synthesis of Class Activation Mapping", is accepted at The 3rd World Conference on eXplainable Artificial Intelligence. Preprint here. Alejandro will present this at XAI in July 2025.
  • Tensorize received a distinguished paper award from CGO. You can find the paper here, and the code here.
December 2024 November 2024
  • Our paper, "Tensorize: Fast Synthesis of Tensor Programs from Legacy Code using Symbolic Tracing, Sketching and Solving", led by Alex Brauckmann, is accepted at CGO 2025.
September 2024 July 2024
  • Slides from my keynote at SYNT 2024 are here.
June 2024
  • A pre-print of our work on Synthetic Programming Elicitation and Repair for Text-to-Code in Very Low Resource Programming Languages is available here.
May 2024
  • I presented our work using program synthesis for program lifting at the VeTSS annual meeting.
April 2024
  • My TedX Talk on Sherlock Holmes and the Case of the Untrustworthy Software is online.
March 2024
  • Our work on Guiding Enumerative Program Synthesis with Large Language Models is accepted at CAV 2024. Yixuan Li is also presenting this work at the EuroProofNet workshop. Pre-print, workshop.
  • The proceedings for AAAI 2024 are now out. Our paper is here.

Talks

Some of the talks I have given are available on YouTube:

Group

Former Group Members