​

Machine Learning

Machine Learning in its most essential form is giving computer systems the ability to visually scan its environment and adapt to

a rapidly changing set of circumstances. Machine learning

technology gives a computer system the ability to see and make rapid decisions based on what is scanned and observed.
Central to this discussion is the importance of adaptation.

The term adaptation is central to biology, particularly evolutionary biology. Adaptation is part of the natural selection process by enabling a person or animal to cope with environmental stresses and pressures. Adaptation, in an AI context, refers to the idea

of a system that is preprogrammed to adapt (make decisions, implement procedures) based on well-defined variables. 

There are two distinguishing features that characterize all machine vision systems. First, the AI system must have the ability to scan
or perceive its surroundings. Second, the AI system must have the ability to evaluate a situation and initiate an appropriate decision/action. As an example, modern aviation relies on an automated flight control management system in order to control the aircraft. The flight control system can control and automate

all phases of a flight operation, including take-off and ascent, flight guidance (autopilot), descent, approach and landing. A second example can be seen with collision avoidance systems on automobiles. For example, a moving vehicle that passes behind
a car when it is pulling out of a driveway will cause the car to stop known as Automatic Emergency Braking (AEB).
 

Autonomous (Self-Driving) Vehicles
One example of real-time adaptation can be seen in the design and functioning of a self-driving vehicle. Companies like Waymo,

Tesla, and Cruise (GM) are engaged in the development of
self-driving vehicles, whereby, the car is designed to do most

of the work of the human driver and the actual person functions more as a passenger. The self-driving car uses sensors that monitor traffic flow, oncoming vehicles, pedestrians, cyclists

and other moving objects. Adaptation and inference is key.

The self-driving car must be able to safely navigate through

constantly changing driving conditions. The car’s intelligence center must be able to know its present geographic location as

well as destination point. The car’s sensors regularly monitor various objects along the roadway according to size, shape and movement pattern. This is where the machine learning part
comes into play. It must be able to differentiate between a moving vehicle, cyclist and pedestrian. The car’s algorithmic software needs to be able to predict and adapt to what these various objects are going to do next.

​