optimal parallel repetition for projection games on low
play

Optimal parallel repetition for projection games on low threshold - PowerPoint PPT Presentation

Optimal parallel repetition for projection games on low threshold rank graphs Madhur Tulsiani, John Wright, Yuan Zhou TTIC CMU CMU Two-prover one-round games Two-prover one-round games Bipartite graph Two-prover one-round games Bipartite


  1. Optimal parallel repetition for projection games on low threshold rank graphs Madhur Tulsiani, John Wright, Yuan Zhou TTIC CMU CMU

  2. Two-prover one-round games

  3. Two-prover one-round games Bipartite graph

  4. Two-prover one-round games Bipartite graph 1. Sample

  5. Two-prover one-round games Bipartite graph 1. Sample 2. Give u to P1 and v to P2

  6. Two-prover one-round games Bipartite graph 1. Sample 2. Give u to P1 and v to P2 3. Receive answers a and b

  7. Two-prover one-round games Bipartite graph 1. Sample 2. Give u to P1 and v to P2 3. Receive answers a and b 4. Accept iff

  8. Two-prover one-round games Bipartite graph 1. Sample 2. Give u to P1 and v to P2 3. Receive answers a and b 4. Accept iff

  9. Two-prover one-round games (biregular) Bipartite graph 1. Sample 2. Give u to P1 and v to P2 3. Receive answers a and b 4. Accept iff

  10. Projection games (biregular) Bipartite graph

  11. Projection games (biregular) Bipartite graph Tests of the form

  12. Projection games (biregular) Bipartite graph Tests of the form is a projection game if for every b , only exists one a to satisfy

  13. Parallel repetition Games

  14. Parallel repetition Games Play as follows:

  15. Parallel repetition Games Play as follows: 1. Sample

  16. Parallel repetition Games Play as follows: 1. Sample 2. Give to P1 and to P2

  17. Parallel repetition Games Play as follows: 1. Sample 2. Give to P1 and to P2 3. Receive answers

  18. Parallel repetition Games Play as follows: 1. Sample 2. Give to P1 and to P2 3. Receive answers 4. Accept iff both and .

  19. Parallel repetition Games Play as follows: 1. Sample 2. Give to P1 and to P2 3. Receive answers 4. Accept iff both and .

  20. Strong parallel repetition is the parallel rep. of times with itself

  21. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ?

  22. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? If , does ?

  23. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? NO!! If , does ?

  24. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? NO!! If , does ?

  25. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? NO!! If , does ? does ?

  26. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? NO!! If , does ? does ? (known as strong parallel repetition )

  27. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? NO!! If , does ? does ? NO!! (known as strong parallel repetition )

  28. Strong parallel repetition is the parallel rep. of times with itself Main Q : How does relate to ? NO!! If , does ? does ? NO!! (known as strong parallel repetition ) no strong parallel repetition in general

  29. Parallel repetition theorems If ,

  30. Parallel repetition theorems If , then [Raz]

  31. Parallel repetition theorems If , then [Raz] if G is a projection game [Rao]

  32. Parallel repetition theorems If , then [Raz] if G is a projection game [Rao] (tight due to the Odd Cycle Game [Raz] )

  33. Parallel repetition theorems If , then [Raz] if G is a projection game [Rao] (tight due to the Odd Cycle Game [Raz] ) no strong parallel repetition for projection games

  34. Expanding games Bipartite graph

  35. Expanding games Bipartite graph G ’s spectrum via the SVD:

  36. Expanding games Bipartite graph G ’s spectrum via the SVD: G is an expanding game if is small.

  37. Expanding games Bipartite graph G ’s spectrum via the SVD: G is an expanding game if is small. G has low threshold rank if is small for some small .

  38. Strong PR for Expanding Games Let be a projection game with 2 nd largest singular value . If , then [Raz and Rosen]

  39. Strong PR for Expanding Games Let be a projection game with 2 nd largest singular value . If , then [Raz and Rosen] Q1: Does this have a tight depencence on ?

  40. Strong PR for Expanding Games Let be a projection game with 2 nd largest singular value . If , then [Raz and Rosen] Q1: Does this have a tight depencence on ? Q2: What about games with low threshold rank ?

  41. Main result Let be a projection game with k -th largest singular value . If , then

  42. Main result Let be a projection game with k -th largest singular value . If , then Improves on Raz and Rosen when k = 2.

  43. Main result Let be a projection game with k -th largest singular value . If , then Improves on Raz and Rosen when k = 2. Optimal for all fixed k .

  44. Proof strategy Parallel repetition framework of [Dinur Steurer] + Strong Cheeger’s inequality due to [KLLGT13]

  45. Proof strategy Parallel repetition framework of [Dinur Steurer] + Strong Cheeger’s inequality due to [KLLGT13]

  46. [Dinur Steurer 2014] New framework for proving parallel repetition theorems using linear algebra .

  47. [Dinur Steurer 2014] New framework for proving parallel repetition theorems using linear algebra . New proof that if G is a projection game [originally from Rao]

  48. [Dinur Steurer 2014] New framework for proving parallel repetition theorems using linear algebra . New proof that if G is a projection game [originally from Rao] Connects PR to Cheeger’s inequality .

  49. Cheeger’s inequality Technique for finding non-expanding sets in graphs.

  50. Cheeger’s inequality Technique for finding non-expanding sets in graphs. Given a graph , let be the second-smallest eigenvalue of its Laplacian. Then

  51. Cheeger’s inequality Technique for finding non-expanding sets in graphs. ( conductance of ) Given a graph , let be the second-smallest eigenvalue of its Laplacian. Then

  52. Cheeger’s inequality Technique for finding non-expanding sets in graphs. Given a graph , let be the second-smallest eigenvalue of its Laplacian. Then

  53. Cheeger’s inequality Technique for finding non-expanding sets in graphs. ( loses a square ) Given a graph , let be the second-smallest eigenvalue of its Laplacian. Then

  54. [Dinur Steurer 2014] Proof that if G is a projection game uses Cheeger’s inequality.

  55. [Dinur Steurer 2014] Proof that if G is a projection game uses Cheeger’s inequality. ( loses a square )

  56. [Dinur Steurer 2014] Proof that if G is a projection game uses Cheeger’s inequality. ( loses a square ) ( because of Cheeger’s inequality)

  57. [Dinur Steurer 2014] Proof that if G is a projection game uses Cheeger’s inequality. ( loses a square ) ( because of Cheeger’s inequality) Better Cheeger’s inequality ⇒ Better PR

  58. [Dinur Steurer 2014] Proof that if G is a projection game uses Cheeger’s inequality. ( loses a square ) ( because of Cheeger’s inequality) Strong Cheeger’s inequality ⇒ Strong PR

  59. Proof strategy Parallel repetition framework of [Dinur Steurer] + Strong Cheeger’s inequality due to [KLLGT13]

  60. Proof strategy Parallel repetition framework of [Dinur Steurer] + Strong Cheeger’s inequality due to [KLLGT13]

  61. [KLLGT13] A strong Cheeger’s inequality for graphs with low threshold rank .

  62. [KLLGT13] A strong Cheeger’s inequality for graphs with low threshold rank . Given a graph , let be the eigenvalues of its Laplacian. Then for every ,

  63. [KLLGT13] A strong Cheeger’s inequality for graphs with low threshold rank . ( when is constant. No square lost!) Given a graph , let be the eigenvalues of its Laplacian. Then for every ,

  64. Main result Let be a projection game with k -th largest singular value . If , then

Recommend


More recommend