1: \begin{abstract}

2: This paper considers nonparametric identification and estimation of the

3: regression function when a covariate is mismeasured. The measurement error

4: need not be classical. Employing the small measurement error approximation,

5: we establish nonparametric identification under weak and easy-to-interpret

6: conditions on the instrumental variable. The paper also provides

7: nonparametric estimators of the regression function and derives their rates

8: of convergence.

9: 

10: \end{abstract}