About This Lesson
It is no secret that computers are getting faster and smaller. Today, you likely have a phone in your pocket that is more powerful than the first computer you ever used. However, this may make you ask whether they are getting smarter or not. Science fiction of the 20th century predicted that we would have flying cars and human-like robots by now. How close are we to achieving that vision?
Artificial Intelligence
Artificial intelligence can be broadly defined as a computer system that can perform cognitive tasks that would normally require a human. This can include thinking, learning, planning, problem-solving, pattern recognition and more.
There are already AI systems that can do all these things to a limited extent. In fact, there are a few systems that would even be able to convince you that they are human, at least for a limited amount of time.
One of the benefits of faster and smaller computers is that scientists have been able to achieve more with them. AI is getting smarter every year. It is also becoming a component of more and more systems including smart business tools and self-driving cars.
Machine Learning
One of the most exciting areas of AI work in recent years has been machine learning. This is an approach to creating artificial intelligence that involves setting up a system that can learn by doing. Scientists create the system then train it with large sets of data.
For example, the creator may want to make a system that can differentiate between images of fruits. The creator would give the system some (a very large amount!) initial data to work with. It would identify patterns and differences between the images and group them accordingly. The creator can give the system target examples that it is trying to find. Alternatively, he or she could simply see what groupings the system can come up with.
AI Vision
Machine learning and AI have helped to create some progress on one of the most exciting technological developments in recent years: self-driving cars. The key to this technology is not in how the car drives itself, but in how it “sees” the world around it.
Using computer vision, the system can identify what parts of an image are cars, street signs, obstacles and more. This is all possible thanks to the developments in machine learning, as well as a concerted effort to teach computers how to better identify patterns in images.
Computer vision is helpful beyond self-driving cars. It can be used in smart surveillance, AI-powered healthcare and more. There is no question, however, that computers are getting smarter; they can finally see!
Evolving AI
Another approach to artificial intelligence is to copy evolution. Scientists create systems that can randomly mutate. The system then selects the result that is “fittest” for its goals. By repeating this over and over again with countless variations, the system can create something that the scientists may not have been able to.
One simple example of this is a team that created a digital character that could walk. Rather than design a character that could walk, the scientists just created a system that could mutate the character and gave it a goal of walking quickly in a straight line. The result was an impressive walking animation created entirely through evolution.
Possible General AI
In the future, it may be possible to create a general AI. This is an artificial intelligence that is equivalent to or even superior to human abilities. In science fiction, this is typically represented as a humanoid robot that can talk, think and act like a super-advanced human. However, general AI does not necessarily need to be contained in a body. It could be an entirely digital solution instead.
Achieving this would be a major breakthrough. However, it is still a little way off.
Learn More
Get started learning more about computers getting smarter. You may have an idea for the next great achievement in AI.
