“…This is in contrast to the usual AdS/CFT correspondence, where due to our limited ability to perform calculations for finite λ in SYM -or equivalently in the full string theory on AdS 5 × S 5 -the relation is a strong/weak coupling duality. Note however, while perturbative calculations in λ of BMN operator two-and three-point functions can be reorganized in λ ′ [14,46,47,49,50] -and hence an extrapolation to large λ seems viable -this is no longer the case for higher point functions: computing for example the 4-point function of tr Z J perturbatively in λ, a naive extrapolation to large λ leads to divergences [51].…”