a union of euclidean
play

A Union of Euclidean Spaces is Euclidean Konstantin Makarychev, - PowerPoint PPT Presentation

A Union of Euclidean Spaces is Euclidean Konstantin Makarychev, Northwestern Yury Makarychev, TTIC AMS Meeting, New York, May 7, 2017 Problem by Assaf Naor Suppose that metric space (, ) is a union of two metric spaces and


  1. A Union of Euclidean Spaces is Euclidean Konstantin Makarychev, Northwestern Yury Makarychev, TTIC AMS Meeting, New York, May 7, 2017

  2. Problem by Assaf Naor Suppose that metric space (π‘Œ, 𝑒) is a union of two metric spaces 𝐡 and 𝐢 that isometrically embed into β„“ 2 . Does π‘Œ necessarily embed into β„“ 2 with a constant distortion? 𝐡 β„“ 2 𝐢

  3. Motivation The problem is closely connected to research in theoretical computer science on β€œlocal -global properties” of metric spaces [Arora, LovΓ‘sz, Newman, Rabani, Rabinovich, Vempala `06; Charikar, M, Makarychev `07] Why are computer scientists interested? Results imply strong lower bounds for Sherali-Adams linear programming relaxations for many combinatorial optimization problems, including Sparsest Cut, Vertex Cover, Max Cut, Unique Games. [Charikar, M, Makarychev `09]

  4. Our Results Q: Suppose that metric space (π‘Œ, 𝑒) is a union of two metric spaces 𝐡 and 𝐢 that embed isometrically into β„“ 2 . Does π‘Œ necessarily embed into β„“ 2 with a constant distortion? A: Yes, π‘Œ embeds into β„“ 2 with distortion at most 8.93. 𝑏 with distortion 𝛽 , 𝐢 β†ͺ β„“ 2 𝑐 with distortion 𝛾 𝐡 β†ͺ β„“ 2 ⇓ 𝑏+𝑐+1 with distortion at most 11𝛽𝛾 π‘Œ = 𝐡 βˆͺ 𝐢 β†ͺ β„“ 2

  5. Approach This talk: consider the isometric case. πœ’ 1 : 𝐡 β†ͺ β„“ 2 πœ’ 2 : 𝐢 β†ͺ β„“ 2 We will define 3 maps: β€’ ΰ΄€ πœ’ 1 : 𝐡 βˆͺ 𝐢 β†ͺ β„“ 2 , a 7-Lipschitz extension of πœ’ 1 to π‘Œ β€’ ΰ΄€ πœ’ 2 : 𝐡 βˆͺ 𝐢 β†ͺ β„“ 2 , a 7-Lipschitz extension of πœ’ 2 to π‘Œ β€’ Ξ” 𝑦 = 𝑒 𝑦, 𝐡 βˆ’ 𝑒(𝑦, 𝐢) πœ” = ΰ΄€ πœ’ 1 βŠ• ΰ΄€ πœ’ 2 βŠ• Ξ”

  6. Approach πœ” = ΰ΄€ πœ’ 1 βŠ• ΰ΄€ πœ’ 2 βŠ• Ξ” Assume that we have β€’ ΰ΄€ πœ’ 1 : 𝐡 βˆͺ 𝐢 β†ͺ β„“ 2 , a 7-Lipschitz extension of πœ’ 1 to π‘Œ β€’ ΰ΄€ πœ’ 2 : 𝐡 βˆͺ 𝐢 β†ͺ β„“ 2 , a 7-Lipschitz extension of πœ’ 2 to π‘Œ β€’ Ξ” 𝑦 = 𝑒 𝑦, 𝐡 βˆ’ 𝑒(𝑦, 𝐢) First, 7 2 + 7 2 + 2 2 πœ” π‘€π‘—π‘ž = πœ’ 1 βŠ• ΰ΄€ ΰ΄€ πœ’ 2 βŠ• Ξ” π‘€π‘—π‘ž ≀ since Ξ” π‘€π‘—π‘ž ≀ 2 .

  7. Approach πœ” = ΰ΄€ πœ’ 1 βŠ• ΰ΄€ πœ’ 2 βŠ• Ξ” β€’ ΰ΄€ πœ’ 1 ensures that distances between points in 𝐡 don’t decrease: πœ’ 1 ȁ 𝐡 = πœ’ 1 is an isometric embedding of 𝐡 into β„“ 2 . ΰ΄€ β€’ ΰ΄€ πœ’ 2 ensures that distances between points in 𝐢 don’t decrease. β€’ Ξ” ensures that distances between points 𝑏 ∈ 𝐡 and 𝑐 ∈ 𝐢 don’t decrease by more than a constant factor.

  8. Approach πœ” = ΰ΄€ πœ’ 1 βŠ• ΰ΄€ πœ’ 2 βŠ• Ξ” 𝑏 𝐡 𝑏′ 𝑐 𝐢 If 𝑒 𝑏, 𝑏′ β‰ͺ 𝑒(𝑏, 𝑐) then πœ’ 2 𝑏 β€² πœ’ 2 (𝑏) βˆ’ ΰ΄€ ΰ΄€ πœ’ 2 (𝑐) β‰ˆ ΰ΄€ βˆ’ ΰ΄€ πœ’ 2 𝑐 = 𝑒 𝑏 β€² , 𝑐 β‰ˆ 𝑒(𝑏, 𝑐) If 𝑒 𝑏, 𝑏′ β‰ˆ 𝑒(𝑏, 𝑐) then β‰₯ 𝑒(𝑏, 𝑏 β€² ) β‰ˆ 𝑒(𝑏, 𝑐) Ξ”(𝑏) βˆ’ Ξ”(𝑐) 𝑏′ is the closest point to 𝑏 in 𝐢

  9. Approach πœ” = ΰ΄€ πœ’ 1 βŠ• ΰ΄€ πœ’ 2 βŠ• Ξ” 𝑏 𝐡 𝑐 𝑏′ 𝐢 If 𝑒 𝑏, 𝑏′ β‰ͺ 𝑒(𝑏, 𝑐) then πœ’ 2 𝑏 β€² πœ’ 2 (𝑏) βˆ’ ΰ΄€ ΰ΄€ πœ’ 2 (𝑐) β‰ˆ ΰ΄€ βˆ’ ΰ΄€ πœ’ 2 𝑐 = 𝑒 𝑏 β€² , 𝑐 β‰ˆ 𝑒(𝑏, 𝑐) If 𝑒 𝑏, 𝑏′ β‰ˆ 𝑒(𝑏, 𝑐) then β‰₯ 𝑒(𝑏, 𝑏 β€² ) β‰ˆ 𝑒(𝑏, 𝑐) Ξ”(𝑏) βˆ’ Ξ”(𝑐)

  10. Constructing maps ΰ΄€ πœ’ 1 and ΰ΄€ πœ’ 2 Goal: Given a map πœ’ ≑ πœ’ 2 : 𝐢 β†’ β„“ 2 find a Lipschitz extension ΰ΄€ πœ’: 𝐡 βˆͺ 𝐢 β†’ β„“ 2 of πœ’ . πœ’ ΰ΄€ 𝐡 β„“ 2 πœ’ 𝐢

  11. Constructing maps ΰ΄€ πœ’ 1 and ΰ΄€ πœ’ 2 Assume that 𝐢 βŠ‚ β„“ 2 and πœ’ = 𝑗𝑒 ; 𝐡 βˆͺ 𝐢 < ∞ . πœ’ ΰ΄€ β„“ 2 𝐡 𝐢

  12. Constructing map ΰ΄€ πœ’ Idea 1: map every 𝑏 to the closest 𝑏 β€² ∈ 𝐢 w.r.t. 𝑒 . Issue: the map may not be Lipschitz. πœ’ ΰ΄€ 𝑏′ 𝑏 𝐡 𝐢

  13. Cover for 𝐡 Let 𝑆 𝑏 = 𝑒 𝑏, 𝐢 for 𝑏 ∈ 𝐡 . 𝐷 βŠ‚ 𝐡 is a cover for 𝐡 if β€’ for every 𝑏 ∈ 𝐡 , there is 𝑑 ∈ 𝐷 s.t. 𝑒 𝑏, 𝑑 ≀ 𝑆 𝑏 and 𝑆 𝑑 ≀ 𝑆 𝑏 β€’ for every 𝑑, 𝑒 ∈ 𝐷 : 𝑒 𝑑, 𝑒 β‰₯ min 𝑆 𝑑 , 𝑆 𝑒 . 𝑆 𝑦 𝑆 𝑦 𝑏 𝑑 𝑑 𝑒 𝑏 ∈ 𝐡 is close to some 𝑑 ∈ 𝐷 points in 𝐷 are β€œseparated”

  14. Cover for 𝐡 Prove by induction that there is always a cover 𝑫 . Let 𝑑 ∈ 𝐡 be the point in 𝐡 with the least value of 𝑆 𝑑 . By induction, there is a cover 𝐷′ for 𝐡 βˆ– Ball 𝑑, 𝑆 𝑑 . Let 𝐷 = 𝐷 β€² βˆͺ {𝑑} . 𝐡

  15. Cover for 𝐡 Prove by induction that there is always a cover 𝑫 . Let 𝑑 ∈ 𝐡 be the point in 𝐡 with the least value of 𝑆 𝑑 . By induction, there is a cover 𝐷′ for 𝐡 βˆ– Ball 𝑑, 𝑆 𝑑 . Let 𝐷 = 𝐷 β€² βˆͺ {𝑑} . 𝑑

  16. Cover for 𝐡 Prove by induction that there is always a cover 𝑫 . Let 𝑑 ∈ 𝐡 be the point in 𝐡 with the least value of 𝑆 𝑑 . By induction, there is a cover 𝐷′ for 𝐡 βˆ– Ball 𝑑, 𝑆 𝑑 . Let 𝐷 = 𝐷 β€² βˆͺ {𝑑} . 𝑑

  17. Cover for 𝐡 Prove by induction that there is always a cover 𝑫 . Let 𝑑 ∈ 𝐡 be the point in 𝐡 with the least value of 𝑆 𝑑 . By induction, there is a cover 𝐷′ for 𝐡 βˆ– Ball 𝑑, 𝑆 𝑑 . Let 𝐷 = 𝐷 β€² βˆͺ {𝑑} . 𝑑

  18. Constructing map ΰ΄€ πœ’ Idea 2: map every 𝑑 ∈ 𝐷 to the closest 𝑑 β€² ∈ 𝐢 . The map is 4-Lipschitz. 𝑔 𝑑′ 𝑑 𝐡 𝐢

  19. Constructing map ΰ΄€ πœ’ Idea 2: map every 𝑑 ∈ 𝐷 to the closest 𝑑 β€² ∈ 𝐢 . The map is 4-Lipschitz. Assume 𝑆 𝑑 ≀ 𝑆 𝑒 . 𝑑′ 𝑑 𝑒 𝑒′ 𝐡 𝐢

  20. Constructing map ΰ΄€ πœ’ Idea 2: map every 𝑑 ∈ 𝐷 to the closest 𝑑 β€² ∈ 𝐢 . The map is 4-Lipschitz. Assume 𝑆 𝑑 ≀ 𝑆 𝑒 . 𝑑′ 𝑑 𝑒 𝑒′ 𝐡 𝐢

  21. Constructing map ΰ΄€ πœ’ Idea 2: map every 𝑑 ∈ 𝐷 to the closest 𝑑 β€² ∈ 𝐢 . The map is 4-Lipschitz. Assume 𝑆 𝑑 ≀ 𝑆 𝑒 . 𝑑′ 𝑑 𝑒 𝑒′ 𝐡 𝐢 𝑒 𝑑 β€² , 𝑒 β€² ≀ 2 𝑒 𝑑, 𝑒 + 2 𝑒 𝑑, 𝑑 β€² ≀ 4𝑒(𝑑, 𝑒)

  22. Kirszbraun Theorem Let 𝐷 βŠ‚ 𝐸 βŠ‚ β„“ 2 and 𝑔 be a Lipschitz map from 𝐷 to β„“ 2 . There exists an extension 𝑕: 𝐸 β†’ β„“ 2 of 𝑔 such 𝑕 π‘€π‘—π‘ž = 𝑔 π‘€π‘—π‘ž 𝐷 β„“ 2 𝐸

  23. Constructing map ΰ΄€ πœ’ Idea 2: map every 𝑑 ∈ 𝐷 to the closest 𝑑 β€² ∈ 𝐢 . Extend 𝑔 from 𝐷 to 𝐡 using the Kirszbraun theorem. 𝑔 𝑑′ 𝑑 𝐡 𝐢

  24. Constructing map ΰ΄€ πœ’ Idea 2: map every 𝑑 ∈ 𝐷 to the closest 𝑑 β€² ∈ 𝐢 . Extend 𝑔 from 𝐷 to 𝐡 using the Kirszbraun theorem. πœ’ 𝑣 = α‰Šπ‘” 𝑣 , if 𝑣 ∈ 𝐡 ΰ΄€ 𝑣, if 𝑣 ∈ 𝐢 πœ’ 𝑣 is 7-Lipschitz: ΰ΄€ β€’ ΰ΄€ πœ’Θ 𝐡 is 4-Lipschitz β€’ ΰ΄€ πœ’Θ 𝐢 is 1-Lipschitz β€’ πœ’ 𝑏 βˆ’ ΰ΄€ ΰ΄€ πœ’ 𝑐 = 𝑔(𝑏) βˆ’ 𝑐 ≀ β‹―

  25. Constructing map ΰ΄€ πœ’ 𝑐 𝑏 𝑔(𝑏) 𝑑 𝑔(𝑑) 𝐡 𝐢

  26. Constructing map ΰ΄€ πœ’ 𝑐 𝑏 𝑔(𝑏) ≀ 𝑆 𝑏 ≀ 4𝑆 𝑏 𝑆 𝑑 ≀ 𝑆 𝑏 𝑑 𝑔(𝑑) 𝐡 𝐢 𝑔 𝑏 βˆ’ 𝑐 ≀ 6𝑆 𝑏 + 𝑒 𝑏, 𝑐 ≀ 7𝑒(𝑏, 𝑐)

  27. Constructing map ΰ΄€ πœ’ 𝑐 𝑏 𝑔(𝑏) ≀ 𝑆 𝑏 ≀ 4𝑆 𝑏 𝑆 𝑑 ≀ 𝑆 𝑏 𝑑 𝑔(𝑑) 𝐡 𝐢 𝑔 𝑏 βˆ’ 𝑐 ≀ 6𝑆 𝑏 + 𝑒 𝑏, 𝑐 ≀ 7𝑒(𝑏, 𝑐) Q.E.D.

  28. Lower Bound There exists a metric space π‘Œ = 𝐡 βˆͺ 𝐢 s.t. β€’ 𝐡 and 𝐢 isometrically embed into β„“ 2 β€’ every embedding of π‘Œ into β„“ 2 has distortion at least 3 βˆ’ 𝜁 π‘œ , where π‘œ = 𝐡 = ȁ𝐢ȁ and 𝜁 π‘œ β†’ 0 as π‘œ β†’ ∞

  29. Open Problems 1. Find the least value of 𝐸 s.t. if 𝐡, 𝐢 β†ͺ β„“ 2 isometrically, then 𝐡 βˆͺ 𝐢 β†ͺ β„“ 2 with distortion at most 𝐸 . We know that 𝐸 ∈ 3, 8.93 . 2. Study the problem for other β„“ π‘ž . We conjecture that the answer is negative for every π‘ž βˆ‰ {2, ∞} . 3. What happens if π‘Œ = 𝐡 1 βˆͺ β‹― βˆͺ 𝐡 𝑙 and each 𝐡 𝑗 β†ͺ β„“ 2 isometrically? We only know that 𝑑 log 𝑙 ≀ 𝐸 ≀ 2 𝐷𝑙 . 4. Assume that every subset of π‘Œ of size Θπ‘ŒΘ isometrically embeds into β„“ 2 . What is the least distortion with which π‘Œ β†ͺ β„“ 2 ? More results and open problems in the paper!

Recommend


More recommend