We provide a filtering algorithm achieving GAC for the conjunction of constraints AtLeast(b,[x[0],x[1],...,x[n-1]],V) /\ Sum(i in 0..n-1)(a[i] c[i]) <= c where the AtLeast constraint enforces at least b variables out of x[0], x[1], ..., x[n-1] to be assigned a value in the set V. This work was motivated by learning simple polynomials, i.e. finding the coefficients of polynomials in several variables from example parameter and function values. We additionally require that coefficients be integers, and that most coefficients be assigned to zero or integers close to 0. These problems occur in the context of learning constraint models from sample solutions of different sizes. Experiments with this more global filtering show an improvement by several orders of magnitude compared to handling the constraints in isolation or with CostGCC, while also out-performing a 0/1 MIP model of the problem.