Optimal Linear Estimation under Unknown Nonlinear Transform

This model is known as the single-index model in statistics, and, among other things, it represents a significant generalization of one-bit compressed sensing.