Research pathwaysComputationPlanned
Formalising published mathematics
Optional research direction: turning published mathematics into machine-checked proofs, using blueprints that link informal statements to formal ones, collaborating on large projects, and contributing to libraries such as Mathlib.
math.comp.formalising-published-mathematics
The lesson for this topic has not been written yet. The outline below is part of the curriculum map and has not been reviewed.
Builds on
Formalising mathematics in Lean and MathlibType theory and propositions as typesReading the mathematical literature
Required prerequisites, all the way down: 22
You will learn to
- Write a blueprint that breaks a published theorem into formalisable lemmas and records their dependencies.
- Formalise a short published result, recording any gaps or implicit assumptions found along the way.
- Describe completed large formalisations and what they did and did not verify.
Definition
- A blueprint is a document linking each informal definition and lemma to its formal counterpart and recording the dependencies between them.
- The trusted base of a formal proof is what must be trusted for the check to be meaningful: the proof checker's kernel, the axioms used and the formal statement.
Theorem
- The Liquid Tensor Experiment formalised a central theorem of Clausen and Scholze's condensed mathematics in Lean; it was completed in 2022.
- The Feit–Thompson odd order theorem was formalised in Coq by Gonthier and collaborators (2012).
- Formalisation verifies the proof of the formal statement; whether that statement faithfully captures the published theorem must be checked by people.
Common misconceptions
- Thinking formalisation makes peer review unnecessary.
- Believing a formal proof follows the informal paper line by line; formal proofs often reorganise the argument substantially.
Optional discoveries
Learning pathways
Further reading
- Jeremy Avigad and Patrick Massot, Mathematics in Lean
- Jeremy Avigad, Leonardo de Moura, Soonho Kong and Sebastian Ullrich, Theorem Proving in Lean 4