Explainable AI: Physics in Machine Learning?

How constraining your ML models with physical principles will make your systems more generalizable and explainable

Imagine you are given a task to predict the number of goals a star footballer is going to make in the next match. Once you got the result, you hurriedly shout out the answer to your manager: minus 3. Within a split second, you realize the improbability of that prediction, and the absurdity of it.