A Reverse Minkowski Theorem
summary
Proves a precise reverse Minkowski: local → sublattice density statement: if every sublattice has det ≥1 then any ball has at most 2^{O(n)} Gaussian mass (and point count), with the worst case morally Z^n and a remaining √log n loss in parameters. Proof converts point counts to the Gaussian mass and Voronoi cell, moves to the compact space of stable lattices, uses projection/induction and convex-geometry results (L-position/B-conjecture inputs). Yields algorithmic and mixing-time consequences (and a 2^{n/2}-time Hermite SVP corollary).
AI-written summary · about
chapters (15)
- 0:00A Reverse Minkowski Theorem
- 4:28Counting Lattice Points: Pop Quiz!
- 5:24Determinant of a Lattice
- 9:02Minkowski's Theorem
- 23:06Applications: Algorithms for Hermite SVP
- 24:03Spoiler
- 24:59Gaussian Mass
- 30:09Space of (Stable) Lattices
- 33:47Ct lies on the boundary
- 36:57Dream Proof Fails
- 41:38Slightly Less Dreamy Proof
- 42:34Our Actual Proof
- 48:29The Actual Theorem
- 50:12Summary
- 51:24Open Questions
related talks
01h 42mMathematics of Lattices
59mSome Recent Results Showing that Lattice Problems Are (sort of) Equivalent in All Norms
01h 08mThe Mathematics of Lattices I
01h 36mAlgorithms for Algebraic Lattices: Classical and Quantum
01h 13mAlgorithms for Lattice Problems
01h 03mComplexity of Lattice Problems
01h 04mThe Mathematics of Lattices II
57mAlgorithms for Lattice Problems, in Practice
