The only way to do great work is to love what you do.
The concept of probably approximately correct (PAC) learning has been foundational in computational learning theory.