After studying and felt in love with machine learning, I think I have succeeded in analytically dealing with my perfectionism. In my case, I was always trying t

After studying and felt in love with machine learning, I think I have succeeded in analytically dealing with my perfectionism. In my case, I was always trying to achieve a local minimum with the cost of 0, and it was overfitting. I was more into integers and algebraic equality rather than floating-point numbers, analytic inequality, probability, and statistics. But after studying machine learning and the related stuff, somehow my mind changed and I now embrace the latter. What is important for me now is to visualize a cost, try to minimize for a general case, stop when passing a threshold, and carefully monitor the whole process. I can plot my current position or cost and see how it change over time and happily stop when things not go well, which I couldn't do in the past that harmed me.