~/talk

Crash Course on Probabilistically Checkable Proofs (PCP): Parallel Repetition

links:none on file
slides:none on file

Parallel repetition takes a projection game and tensorizes it so questions are d-tuples; the basic PCP route uses this to amplify a 1 versus 1−ε gap to 1 versus ≈0. Raz proves exponential decay of value but also shows strong parallel repetition is false: the odd-cycle counterexample has single-game value 1−1/m yet repeated value ≈(1−1/m^2)^d, not (1−1/m)^d. SDP relaxations explain the gap, and the Raz-based constructions produce tilings of R^d with surface area ~√d, so the counterexamples and geometric foams work.

AI-written summary · about

  1. 0:00Introduction to PCPs
  2. 1:04Basic PCP theorem
  3. 3:05Tightness and 3-SAT
  4. 6:00From Label Cover to Strong PCP
  5. 7:55Label cover definition
  6. 12:12Hardness of label cover
  7. 15:12Two-player game view
  8. 18:48Parallel repetition
  9. 23:32Parallel repetition theorem
  10. 27:35Tightness in repetition
  11. 30:27Unique games and max-cut
  12. 34:16Unique games conjecture
  13. 37:28Strong parallel repetition
  14. 42:39Odd cycle game
  15. 45:55Geometric interpretation
  16. 58:42Necessity of the two
  17. 1:02:28Questions and discussion
~/search
type to search 6,249 conferences, 138,987 speakers and 185,172 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.