r/MachineLearning - [D] Can we minimize counting cost function for perceptron algorithm?

The question is posted on Cross Validated (haven't figured out how to write formulas in reddit). Feel free to leave your comments either here or at Cross Validated.