Review # 1

We thank the reviewers for their careful reading of the paper. Please find the answers to the questions below. But that being said, it's important to Q: I wonder if you could use these arguments to show guarantees based on the perceptron algorithm itself. A: The reviewer's suggestion is correct; a slight modification of the perceptron algorithm, where an update is performed We will add more detail about this in the revised version. Q: The result is stated for a "small constant" A: We view simplicity as an advantage of our work.