Google's slew of big announcements in recent weeks — unveiling Google Translate's neat new machine learning tricks, its voice-controlled speaker system Home, and its new smartphone line Pixel (itself equipped with the futuristic-feeling virtual Assistant) — are evidence enough that the tech giant is investing heavily in artificial intelligence. But this Fortune story suggests that's just the tip of the iceberg: Google has dramatically ramped up its investments in what's known as "deep learning" in the last four years. In 2012, Google had two deep-learning projects underway. Today, Google has more than 1,000 deep-learning projects in every major product category, including search, maps, translation, and self-driving cars.
Deep learning refers to a kind of AI that enables software to train itself to perform tasks, like speech and image recognition. The algorithms behind this kind of AI have been around since the 1980s and 1990s, Fortune reports, but it's only in recent years that technology companies have been able to harness its true potential thanks to the explosion of easily accessible big data. Google launched its deep-learning project Google Brain in 2011, and in 2014 it bought DeepMind, a pioneering reinforcement learning program that's been shattering AI landmarks in gameplay this year.