|
Raw computational speed is not where the most important developments will come from. The big game to be won is in the ability of computers to integrate input from the new emerging sensors and to be able to learn what to do with this data on their own. Right now huge volumes of sensor data have to be first managed by humans and then put into forms the computers can do calculations on to make decisions. "Deep Learning" so that the new eyes and ears of the computer (new sensors of all kinds , not just audio and video) can communicate directly with the brain (computational speed and multilevel processing software) to eliminate the need for humans to be the interface between the sensors and the brain.
|