A FORMAL PROOF OF THE KEPLER CONJECTURE
By T. Hales, Mark Adams, G. Bauer, D. T. Dang, John Harrison, T. Hoang, Cezary Kaliszyk, Victor Magron, Sean McLaughlin, T. Nguyen and 12 more. Forum of Mathematics, Pi, 2015
523 citations recorded by the source corpus.
This article describes a formal proof of the Kepler conjecture on dense sphere packings in a combination of the HOL Light and Isabelle proof assistants. This paper constitutes the official published account of the now completed Flyspeck project.
Subfields: Geometry and Topology, Statistical and Nonlinear Physics, Theoretical Computer Science
View Paper: https://www.semanticscholar.org/paper/de5eff4b6c605fa7318adfef99f5705416153a7b
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Geometry and Topology
- Stewart Wilcox
- Pere-Pau Vázquez
- Percy Liang
- Jia Deng
- Timo Ropinski
- Tobias Ritschel
- Josef Urban
- Abhinav Kumar
In Statistical and Nonlinear Physics
- Alexander A. Alemi
- Pere-Pau Vázquez
- Percy Liang
- Jia Deng
- Timo Ropinski
- Tobias Ritschel
- Markus Norman Rabe
- Guillaume Lample
In Theoretical Computer Science
- Marie-Anne Lachaux
- Percy Liang
- Lawrence Charles Paulson
- Guillaume Lample
- Josef Urban
- Lei Yu
- Wenda Li
- Yuhuai Wu
Citations in the corpus
- A conceptual breakthrough in sphere packing (2016, background): “The three-dimensional problem was solved by Hales [9] via a lengthy and complex computer-assisted proof, which was extraordinarily difficult to check but has since been completely verified using formal logic [10].”
- Learning to Prove Theorems via Interacting with Proof Assistants (2019, methodology): “HOList consists of 29K proofs solely from the formalization of the Kepler conjecture (Hales et al., 2017), a theorem in discrete geometry, whereas CoqGym covers more diverse domains including not only pure mathematics but also computer systems.”
- Monte Carlo convolution for learning on non-uniformly sampled point clouds (2018, background): “The Kepler conjecture [Hales et al. 2017], provides an upper bound to the number of points inside the receptive fields: This is illustrated in Fig.”
- HOList: An Environment for Machine Learning of Higher-Order Theorem Proving (extended version) (2019, background): “Our choice fell on HOL Light, the interactive theorem prover (ITP) in which the proof of the Kepler conjecture [7] has been formalized.”
- Developing Bug-Free Machine Learning Systems With Formal Mathematics (2017, background): “…Chli-pala, 2013; Chen et al., 2015), which has aimed to prove properties of computer programs, and formal mathematics (Rudnicki, 1992; Gonthier, 2008; Gonthier et al., 2013; Hales et al., 2015), which has aimed to formally represent and generate machine-checkable proofs of mathematical theorems.”
- HyperTree Proof Search for Neural Theorem Proving (2022, background): “Second, formalizing an existing proof still involves significant effort and expertise (the formalization of the Kepler conjecture took over 20 person years to complete [3]) and even seemingly simple statements sometimes remain frustratingly challenging to formalize.”
- DeepMath - Deep Sequence Models for Premise Selection (2016, methodology): “Examples in mathematics include the HOL Light proof of the Kepler conjecture (Flyspeck project) [14], the Coq proofs of the Feit-Thompson theorem [12] and Four Color theorem [11], and the verification of most of the Compendium of Continuous Lattices in Mizar [2].”
- Modelling High-Level Mathematical Reasoning in Mechanised Declarative Proofs (2020, background): “Over the last decade, nontrivial mathematical theorems [16, 18, 19], security protocols [14, 40], and industry-scale software systems [27, 32] have been mechanically checked, highlighting numerous issues and bugs.”
- Improving Graph Neural Network Representations of Logical Formulae with Subgraph Pooling (2019, background): “cial intelligence research that has demonstrated utility in the design of operating systems [1, 2], distributed systems [3, 4], compilers [5, 6], microprocessor design [7], and in general mathematics [8]. Classical automated theorem provers (ATPs) have historically been most useful for solving probl”
- The sphere packing problem in dimension 24 (2016, background): “Aside from the trivial case of one dimension, the optimal density was previously known only in two [11], three [7], [8], and eight [12] dimensions, with the latter result being a recent breakthrough due to Viazovska; see [1], [9] for expositions.”
- Premise Selection for Theorem Proving by Deep Graph Embedding (2017, unknown): “This limitation gave rise to the development of interactive theorem proving [11] such as Coq [14] and Isabelle [15], which combines humans and machines in theorem proving and has led to impressive achievements such as the proof of the Kepler conjecture [16] and the formal proof of the Feit-Thompson ”