We prove that almost all random subsets of a finite vector space are weak Salem sets (small Fourier coefficient), which extend a result of Hayes to a different probability model.

We improve the estimates in the restriction problem in dimension n⩾4. To do so, we establish a weak version of a k-linear restriction estimate for any k. The exponents in this weak k-linear estimate are sharp for all k and n.

We prove the boundedness of a class of tri-linear operators consisting of a quasi piece of bilinear Hilbert transform whose scale equals to or dominates the scale of its linear counter part. Such type of operators is motivated by the tri-linear Hilbert transform and its curved versions.