HepLean: Digitalising high energy physics

Joseph Tooby-Smith

2024-05-15

Abstract:We introduce HepLean, an open-source project to digitalise definitions, theorems, proofs, and calculations in high energy physics using the interactive theorem prover Lean 4. HepLean has the potential to benefit the high energy physics community in four ways: making it easier to find existing results, allowing the creation of new results using artificial intelligence and automated methods, allowing easy review of papers for mathematical correctness, and providing new ways to teach high energy physics. We will discuss these in detail. We will also demonstrate the digitalisation of three areas of high energy physics in HepLean: Cabibbo-Kobayashi-Maskawa matrices in flavour physics, local anomaly cancellation, and Higgs physics.

High Energy Physics - Phenomenology,Logic in Computer Science,High Energy Physics - Theory

What problem does this paper attempt to address?

This paper introduces an open-source project called HepLean, which aims to digitize definitions, theorems, proofs, and calculations in high energy physics using the interactive theorem prover Lean 4. The main goal of this project is to make results more easily accessible, allow for AI and automation methods to create new findings, simplify mathematical correctness verification for papers, and provide new ways of teaching. HepLean leverages the MathLib library, which has already digitized mathematical results including representation theory and manifold theory, enabling physicists to express high energy physics results in Lean without directly dealing with dependent type theory. The paper also demonstrates how concepts in fields like CKM matrix, local anomaly cancellation, and Higgs physics can be represented in Lean code.

With HepLean, authors can simplify the proof process for new theorems, utilize automated proof strategies to expedite calculations, and ensure mathematical accuracy. Additionally, it provides new avenues for teaching, such as gamified learning and project-based digitized learning. Currently, HepLean has made progress in flavor physics with the Cabibbo-Kobayashi-Maskawa matrix, local anomaly cancellation, and Higgs physics.

The motivation behind this paper is to address the difficulty of finding results scattered across the literature in high energy physics, improve efficiency in creating new findings, simplify mathematical correctness verification, and provide new teaching tools. Through these means, HepLean has the potential to make a positive impact on the high energy physics community.