![]() *Commanders Intent, Empowerment, Ethics and Morals* That may be cool, yes, but what is the point? That's what we'll talk about next time. Now, given the position and speed of the ball about 5-10 feet from where it's released, it can at times predict where the ball will be within about 4.8 cm, or 1.8 inches. ![]() ![]() That's right, I learned neural network programming because I'm gloriously lazy. The physics of projectile motion is well established, and given the same inputs I've given the network, you could calculate the same thing with a high degree of accuracy.īut given the choice, I'm going to set up a network that I can automate a whole heckin bunch of randomized throws and let the computer learn the math for me. If this seems simple.well, you're not wrong. In my case, what I'm doing is feeding the network the information about a ball that is thrown shortly after it's released, and using the network to predict where it will cross a plane 6m (almost 20 feet) away. I understand what it's doing-essentially sending inputs through a matrix that predicts likely outcomes-but it still is amazing seeing it learn to hone down to a more and more accurate prediction. So having finally gotten around to writing and training my own little neural network, it still strikes me as some kind of magic. I did some projects with Python and scikit-learn, but I'm always motivated to start at the lowest level possible, so I can understand on a conceptual level what's happening. I remember looking at state machine diagrams and fuzzy logic algorithms alongside my computer networking assignments in class back in 2003.īut it took me a long time to wrap my head around the kind of right-brain big picture thinking needed to get into machine learning properly. I've always been interested in AI as it pertains to gaming.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |