Machine Learning.


Earlier this year, Facebook's Mark Zuckerberg, Tesla Motor's Elon Musk, and several other high profile investors capitalized on a secretive artificial intelligence lab named "Vicarious". As a private company, details of its business remained appropriately classified, but sources noted that they were in the artificial neural networks arena with research focused on software emulating the human neocortex. They were a machine learning company.

What is Machine Learning? A quick drop in on Wikipedia will tell you that it is defined as "a branch of artificial intelligence, [concerned] with the construction and study of systems that can learn from data". Artificial intelligence has, at large, remained in the realm of science fiction novels and movies, the general public often scoffing its premise as nonsense and a study without realistic scientific substance. This has been true in the past, but academics and private enterprises alike are now realizing the increasing power of computing technologies and its application directed toward the development of intelligent machines.



Machine learning theory (also known as Cognitive Computing so named in the academic environment) has several topics to note: supervised learning, algorithms computing on data in a controlled lab environment, inputs and outputs are generally controlled; unsupervised learning letting loose a machine on unstructured polluted data with general classifiers to organize the data and produce predictable and knowledgable results, or reinforced learning which treats a computer like a child - reward system in hand. Do something good, you get a reward, do something bad, you are punished.

Several approaches have gained momentum as of late, for example, Google's foray into machine learning has resulted in strong interest in Artificial Neural Networks, or the emulation of brain circuitry, hierarchy of learning, led by it's head of research Geoff Hinton (sometimes known as the father of artificial neural networks). If you've ever used Google Now's features, for example, asking it "What is the weather like" or "Find me images of cats". Chances are you have interacted with early forms of the algorithm in the internet search engine.

I love talking about this stuff because I strongly believe in its future applications. Give a  robot a Google brain (connected to the cloud) and you have a surprisingly useful and sophisticated mechanical apparatus that learns independently. Our robots assistants are on their way.

Comments

Popular posts from this blog

My Future Career.

Turkey. A Travel Log.

Corporate restructuring.