For many people, the idea of artificial intelligence is the stuff of science fiction. They think of android on spaceships vastly outperforming humans in terms of math and analytical thinking but lagging in creativity and emotion. However, AI isn’t just limited to sci-fi. In fact, it is very real and advances in machine learning are helping to bring out favorite sci-fi novels, movies and TV shows closer to reality.
1) Improved Natural Language Capabilities
One key area of AI development has been computers that can understand and converse with people. Chatbots are an example of this. You have likely already experienced one when interacting with a customer service department. They are growing increasingly popular because they can help customers with no wait times.
Similar technology has developed with homes speakers and phone assistants. These tools can interpret what a person is saying and figure out an appropriate response. This iteration is even more impressive because it also has the step of parsing the audio signal of the person speaking. With accents and different dialects, this can be challenging.
While many of these systems are somewhat limited in scope (you can mostly just ask certain commands or get search results), they are improving rapidly. More significantly, the systems are better at understanding different iterations of commands. For example, you could ask who is John Doe or tell me about John Doe. Machine learning has helped these systems to understand what you mean even though the wording is different.
2) Computers That Can See
Computers are also learning how to see things. Not long ago, images were the gold standard of figuring out whether a person was real or a computer program (captcha). Today, thanks to computer vision technology, computer systems can identify objects in images they are receiving.
This can be done in real time using a video camera. It is at the heart of many modern self-driving vehicle technologies. If you have a car that can stay in a lane on the highway or slow down when a car in front of it stops, your vehicle is likely using computer vision (possibly in combination with radar sensors).
By interpreting patterns in an image, the system can find objects. Machine learning lets the system sort those images into groups. For example, it may know what stop signs look like. It can also determine what cars look like. By using a huge number of similar inputs from previous images, the system can determine exactly what the object is and where it is relative to the camera.
3) Complex Task Completion
Another element of artificial intelligence is complex problem-solving. Not long ago, an AI that could play chess at a competitive level was serious news. Today, artificial intelligence is capable of playing games that are way more complicated than chess.
In chess, each player controls at most 16 pieces, all of which can make a limited number of possible moves. Only one piece moves per turn. So, the computer is making a choice from a relatively limited number of possible options.
Today, computers are completing three-dimensional games with an exponentially greater number of options. This is achieved through machine learning. The system can collect extensive data on the results of different inputs and make relevant choices.
While playing games with AI may seem relatively unimportant, it is worth noting that play is exactly how humans learn to solve problems when we are young. Furthermore, humans are capable of solving very complex problems, in part, because we only evaluate the reasonable options. Today, computers are not only solving low-stakes problems in a manner similar to humans, they are often exceeding our capabilities.
Discover More About Machine Learning
Learn more about how machine learning is advancing today. You may be impressed by how intelligent computer systems are becoming. With leaps and bounds, this area is moving forward. It will have a lot of important implications for business and everyday life.
