Go to the profile of Damian Ali
Damian Ali

Paving the way for AI

Most people seem to think “AI” or artificial intelligence is a thing of the distant future. But we can’t be further from the truth. While we may not be at the “have your own robot to perform simple household tasks” days just yet. AI and Deep Learning are being utilized to predict what we want to eat, watch, and buy. As well as facial recognition and, now, self driving cars are right around the corner.

In January, the US Department of Transport said it would be willing to waive some regulations to get more self-driving cars onto the roads.

Late last year, Google submitted their design for a self-driving car to the US National Highway Traffic Safety Administration (NHTSA) outlining the design. Earlier this month, the NHTSA responded agreeing with Google, explaining the software could qualify as the driver for the vehicle.

Nvidia utilizes Deep Learning coupled with cameras and sensors to absorb information around a vehicle, which is then fed to their NVIDIA DRIVE auto-pilot car computer.

An example of deep learning, training a self-driving car to recognize different classes of vehicles so it can take action appropriate to each.

An example of Deep Learning or Machine Learning would be Nvidia engineers recording around 40 hours of video from the cameras mounted on their cars, which was then manually tagged by frame. Taking the 68000 or so objects in the footage, the engineers fed these images to servers, which began to study them by identifying their patterns, shapes, and angles. Over time, the computer starts to identify object, such as a car. However, Deep Learning allows the computer to categorize the information in layers. These layers are then stacked and the machine learns what computer scientists call a “hierarchical representation.” This is where each layer looks for a specific shape or pattern and passes it up, allowing it to eventually learn and understand the concept of a person crossing the street, for instance.

What’s next? Further down the road, I see Deep Learning as a crucial tool towards each and every one of us with our own personal robot helper. Perhaps, giving each and every piece of technology we own the ability to know what we want, and when we want it. Question is, whether or not what we’re building for ourselves is bringing us together or pushing us further apart?