The dream of science fiction, creation of a purely independent digital mind has been a pursuit of companies like Amazon, Apple, Google, & Microsoft for over half a decade. With their dependency growing ever-faster. Their capability boundless. Tech giants are looking for the next batch of deep neural programmers to lead the edge. Creating applications such as AlphaGo & AlphaGo Zero we have a program that shows us what the future holds.
An allgorithim that has the ability to learn, this is a long standing goal in artificial intelligence that a program can learn with superhuman proficiency in challenging disciplines. In 2015, AlphaGo became the first A.I. to defeat a world champion at a game of GO, without any handicaps. The tree search in AlphaGo calculated positions and picked moves using deep neural networks. Which were learned through watching human experts own playing ability, and strengthened through learning from playing itself.
Two years later in 2017 AlphaGo Zero was released with an allgorithim solely based on self-learning abilities. Without human data, instruction, or realm understanding beyond the game rules. AlphaGo Zero became its own teacher with a neural network trained to predict AlphaGo's moves as well as the winner of AlphaGo's games. This neural network improves the capabillites of the tree search, with greater results from move selection and stronger self-play in the following iteration. Starting tablua rasa, AlphaGo Zero achieved superhuman results, winning 100-0 against AlphaGo.
Our current students and future programmers will be expected to solve problems we nor they can conceive of yet. The coding languages of the future will be vast and growing from what they are today. Let us teach them the ability to solve their problems with their own creativity and problem solving. Creating what was once the dream of science fiction.