Mathematical Knowledge Base Development and Formalization
Develop a knowledge base for covering groups, with AI-assisted mathematical exposition and Lean-checked proofs.
Task 1 concerns the structural foundations of covering groups, while Tasks 2 and 3 concern character theory and Shimura liftings, using harmonic-analytic and geometric tools. This selection is inevitably shaped by our interests and represents only a small part of the literature on the Langlands program for covering groups. We hope it will stimulate further work in this direction.
Tasks are listed in decreasing order of priority.
Task 1: Brylinski–Deligne, Weissman and Yifei Zhao — Covers of general reductive groups and the principal-block local Langlands correspondence
Formalize Brylinski–Deligne’s classification of central K2-extensions, Weissman’s L-groups, and Yifei Zhao’s construction and classification of étale metaplectic covers and their dual data.
The principal-block part concerns the established Iwahori-spherical correspondence and its Hecke-algebra description.
References
- Brylinski–Deligne, Central extensions of reductive groups by K2, Publ. Math. IHES 94 (2001), 5–85. DOI.
- Weissman, L-groups and parameters for covering groups, Astérisque 398 (2018), 33–186. DOI.
- Yifei Zhao, Étale metaplectic covers of reductive group schemes, J. Eur. Math. Soc. (2025), published online first. DOI; arXiv.
Further references
- Luozi Shi and Yifei Zhao, What are the extended pure inner forms of a cover?, preprint (2025). arXiv.
- Gordan Savin, On unramified representations of covering groups, J. reine angew. Math. 566 (2004), 111–134. DOI.
- McNamara, Principal Series Representations of Metaplectic Groups Over Local Fields, in Multiple Dirichlet Series, L-functions and Automorphic Forms, Progress in Mathematics 300 (2012), 299–327. DOI.
Task 2: Kazhdan–Patterson — Representations, automorphic forms and metaplectic correspondence for covers of general linear groups
Formalize the local and global theory of Kazhdan–Patterson covers of GLr: metaplectic forms, the classification of irreducible representations by Kaplan–Lapid–Zou, and the metaplectic correspondence of Flicker and Flicker–Kazhdan via character identities and trace-formula comparisons.
References
- Kazhdan-Patterson, Metaplectic forms, IHES 59 (1984), 35-142. DOI.
- Kazhdan-Patterson, Corrections to Metaplectic forms, IHES 62 (1985), 203. DOI.
- Kaplan-Lapid-Zou, Classification of irreducible representations of metaplectic covers of the general linear group over a non-archimedean local field, Representation Theory 27 (2023), 1041-1087. DOI.
- Flicker, Automorphic forms on covering groups of GL(2), Invent. Math. 57 (1980), 119-182. DOI.
- Kazhdan-Patterson, Towards a generalized Shimura correspondence, Adv. Math. 60 (1986), 161-234. DOI.
- Flicker-Kazhdan, Metaplectic correspondence, IHES 64 (1986), 53-110. DOI.
Task 3: Renard–Trapa and Adams–Barbasch–Paul–Trapa–Vogan — Vogan duality, Kazhdan–Lusztig–Vogan algorithms and unitary Shimura correspondences for real covering groups
Formalize David Renard and Peter Trapa’s Kazhdan–Lusztig–Vogan character algorithms and Vogan character-multiplicity duality for the real covering groups treated in their papers, including metaplectic groups and double covers of real general linear and unitary groups.
Formalize the unitary Shimura correspondence for split real groups developed by Adams, Barbasch, Paul, Trapa and Vogan through harmonic-analytic methods.
The proved results include the complementary-series injection for the stated classical groups and the unitarity equivalence under Howe correspondence for the metaplectic symplectic representations containing the det^(1/2) K-type. The broader correspondence is stated as a conjecture in that work.
References
- David Renard and Peter Trapa, Irreducible genuine characters of the metaplectic group: Kazhdan–Lusztig algorithm and Vogan duality, Representation Theory 4 (2000), 245–295. DOI.
- David Renard and Peter Trapa, Irreducible characters of the metaplectic group II: Functoriality, J. reine angew. Math. 557 (2003), 121–158. DOI.
- David Renard and Peter Trapa, Kazhdan–Lusztig algorithms for nonlinear groups and applications to Kazhdan–Patterson lifting, American Journal of Mathematics 127 (2005), 911–971. DOI.
- Adams-Barbasch-Paul-Trapa-Vogan, Unitary Shimura correspondences for split real groups, JAMS 20 (2007), 701-751. DOI.
Further references
Knowledge-base requirements
For a sample knowledge base, see our blueprint for the combinatorial invariance of Kazhdan–Lusztig polynomials, developed from OpenAI’s work. This serves as a baseline example; submitted knowledge bases are expected to surpass it in quality.
Main theorems and their proofs. State the main theorems with explicit hypotheses, conclusions and source references. Provide a readable mathematical proof and a checked Lean proof of each theorem, and verify that the written and formal statements agree. Encode definitions directly, take chosen objects as input data, and prove the well-definedness and properties of constructions where required. Link the Lean declarations and record the checked revision, dependencies and axioms. Retain the assumptions of conditional results and distinguish them from conjectures.
Generality. State definitions and results in the greatest generality supported by the arguments. Keep parameters explicit, identify where each assumption enters the proof, and derive important special cases from the general results.
Consistent notation and conventions. Use a consistent set of notation and conventions throughout the exposition and Lean code. Explain how alternative formulations relate, including the equivalences or comparison maps needed to move between them.
Reusable Lean interfaces and extensible type design. Build one coherent library on mathlib's types, structures and theorems, following its notation and API conventions. Choose general data structures and reusable interfaces that keep routine constructions and theorem applications simple. Use structures, typeclasses and explicit parameters according to their roles, keeping instance inference predictable and coercions coherent. Make multiple inheritance paths to the same structure agree definitionally to avoid conflicting instance diamonds. Check the design on representative examples and extensions.
Explanations, examples and applications. Explain what the main definitions capture. Introduce the idea of a proof before its details and work out significant calculations. Include worked examples and applications showing how the results can be used in further arguments.
Linked concepts and layered navigation. Make every mathematical concept in a statement clickable, with symbols and hypotheses linked to the definitions and conventions used there. Let readers move from a short explanation to full definitions, examples and prerequisite material. Provide direct links to proofs and Lean code. Show each page's place in the topic, its connections to other results and a clear route back to the original statement.
Attribution, historical context and references. For important theorems, name the people who proved them and explain their place in the development of the subject, including later generalizations and corrections. Give full references with authors, titles, publication details, page or theorem locations, editions and DOI links. Identify contributors to the exposition and Lean formalization as well.