Download presentation
Presentation is loading. Please wait.
Published byTrinity Asch Modified over 9 years ago
1
Discrepancy Minimization by Walking on the Edges Raghu Meka (IAS/DIMACS) Shachar Lovett (IAS)
2
1 2 3 4 5 Discrepancy 1*11* *11*1 11111 ***11 1*1*1 1 2 3 4 5 1*11* *11*1 11111 ***11 1*1*1 3 1 1 0 1
3
Discrepancy Examples Fundamental combinatorial concept Arithmetic Progressions
4
Discrepancy Examples Fundamental combinatorial concept Halfspaces Alexander 90: Matousek 95:
5
Discrepancy Examples Fundamental combinatorial concept Axis-aligned boxes Beck 81: Srinivasan 97:
6
Why Discrepancy? Complexity theory Communication Complexity Computational Geometry Pseudorandomness Many more!
7
Spencer’s Six Sigma Theorem Central result in discrepancy theory. Beats random: Tight: Hadamard. Spencer 85: System with n sets has discrepancy at most. “Six standard deviations suffice”
8
Conjecture (Alon, Spencer): No efficient algorithm can find one. Bansal 10: Can efficiently get discrepancy. A Conjecture and a Disproof Non-constructive pigeon-hole proof Spencer 85: System with n sets has discrepancy at most.
9
This Work Truly constructive Algorithmic partial coloring lemma Extends to other settings New elemantary constructive proof of Spencer’s result EDGE-WALK: New algorithmic tool
10
Outline 1.Partial coloring Method 2.EDGE-WALK: Geometric picture
11
Focus on m = n case. Lemma: Can do this in randomized time. Partial Coloring Method Input: Output:
12
Outline 1.Partial coloring Method 2.EDGE-WALK: Geometric picture
13
1*11* *11*1 11111 ***11 1*1*1 Discrepancy: Geometric View 1 1 1 3 1 1 0 1 3 1 1 0 1 1 2 3 4 5
14
1*11* *11*1 11111 ***11 1*1*1 Discrepancy: Geometric View 1 1 1 3 1 1 0 1 1 2 3 4 5
15
Discrepancy: Geometric View Goal: Find non-zero lattice points in Polytope view used earlier by Gluskin’ 88.
16
Claim: Will find good partial coloring. Edge-Walk Start at origin Gaussian walk until you hit a face Gaussian walk within the face Goal: Find non-zero lattice point in
17
Edge-Walk: Algorithm Gaussian random walk in subspaces Standard normal in V: Orthonormal basis change
18
Edge-Walk Algorithm Discretization issues: hitting faces Might not hit face Slack: face hit if close to it.
19
Edge-Walk: Algorithm
20
Edge-Walk: Intuition 1 100 Hit cube more often! Discrepancy faces much farther than cube’s
21
Summary 1.Edge-Walk: Algorithmic partial coloring lemma 2.Recurse on unfixed variables Spencer’s Theorem
22
Open Problems Q: Other applications? General IP’s, Minkowski’s theorem? Some promise: our PCL “stronger” than Beck’s Q:
23
Thank you
24
Main Partial Coloring Lemma Algorithmic partial coloring lemma
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.