Machine learning, commonly abbreviated "ML," is a type of artificial intelligence (AI) that "learns" or adapts over time. Instead of following static rules coded in a program, ML technology identifies input patterns and contains algorithms that evolve over time.
Machine learning has a wide variety of applications, many of which are now part of everyday life. Below are a few examples:
- Medical diagnoses
- Autonomous vehicles
- Online ad targeting – Google AdSense and Facebook Advertising
- Speech recognition – Google Assistant, Amazon Alexa, Microsoft Cortana, and Apple Siri
- Image recognition – Google image search, facial recognition on Facebook and in Apple Photos
Self-Driving Vehicle Example
Autonomous vehicles incorporate machine learning to improve their safety and reliability. A self-driving car that uses traditional artificial intelligence can respond to any road conditions it has been programmed to handle. However, if the software encounters unrecognized input, the car may default to a backup safety measure, such as slowing down, stopping, or requiring a manual override.
Machine learning can enable a vehicle to recognize events and objects that have not been explicitly programmed in the source code. For example, a car may be programmed to recognize street lights, but not flashing lights on construction barricades. By learning from experience — possibly recording the driving behavior of a human driver — the car will start to recognize construction barriers and respond accordingly.
ML technology is what enables autonomous vehicles to differentiate between objects on the road, such as cars, bikes, humans, and animals. It also helps automobiles drive more reliably in imperfect weather conditions and on roads without clear lines. The goal is to enable vehicles to drive like humans while avoiding mistakes caused by human error.
Updated: June 10, 2019