SQ Lower Bounds for Learning Single Neurons with Massart Noise

Neural Information Processing Systems 

We study the problem of PAC learning a single neuron in the presence of Massart noise.