Artificial intelligence is This. It is overblown, incompletely understood, and flawed but is already core to our lives—and it will only broaden its reach.
AI power driverless car researchdifference Unseen signs of illness on medical imaging, finds answers when you ask Alexa questions, and lets you unlock your phone with your face to talk to your friends as animated stools on your iPhone X using Apple’s Animoji. Those are just some of the ways AI has touched our lives, and there is still a lot of work to be done. But don’t worry, super smart algorithms won’t get all the jobs or wipe out humanity.
The current boom in every field of AI is catalyzed by breakthroughs in an area known as machine learning. It involves “training” the computer to perform tasks based on examples, rather than relying on human programming. A technique called deep learning has made this method much more powerful. Just ask Lee Sedol, who holds 18 international titles in the intricate game of Go. He there is ice cream with software called AlphaGo in 2016.
There is evidence that AI can make us happier And healthier. But there is also reason to be cautious. Incidents where algorithms pick up on or amplify surrounding social biases species or sex shows that an AI-enhanced future will not automatically become a better future.
The beginning of artificial intelligence
Artificial intelligence as we know it started as a resort project. Professor John McCarthy of Dartmouth College coined the term in the summer of 1956, when he invited a small group to spend a few weeks thinking about how to make machines do things like use language.
He has high hopes for a breakthrough towards human-level machines. “We think a significant step forward can be made,” he wrote with his co-hosts“if a carefully selected group of scientists come together to study it over a summer.”
Those hopes were not fulfilled, and McCarthy then admit it that he was too optimistic. But the panel helped researchers dream of intelligent machines merge into a recognized academic field.
Early work often focused on solving rather abstract problems in mathematics and logic. But it wasn’t long before AI began to show promising results in more human tasks. In the late 1950s, Arthur Samuel created programs to learn to play checkers. In 1962, a man scored a victory over a master in the game. In 1967, a program called Dendral showed it could replicate the way chemists interpret mass spectrometry data about the structure of chemical samples.
As the field of AI evolves, so do different strategies for creating smarter machines. Some researchers have tried to distill human knowledge into code or make rules for specific tasks, such as understanding language. Others are inspired by the importance of learning to understand human and animal intelligence. They have built systems that can better accomplish a task over time, perhaps by simulating evolution or by learning from example data. The field hit a major milestone after computers mastered tasks that were previously only able to be completed by humans.
Deep learning, the rocket fuel of the current AI boom, is a revival of one of the oldest ideas in AI. This technique involves transmitting data through mathematical networks inspired by the activity of brain cells known as artificial neural networks. As a network processes training data, the connections between parts of the network adjust, building the ability to interpret future data.
Artificial Neural Networks became an age-old idea in AI not long after the Dartmouth conference. For example, the room-filling Perceptron Mark 1 since 1958 has learned to distinguish different geometric shapes and has write on IN The New York Times “The embryo of a computer designed to read and grow wiser”. But neural networks fell out of favor after an influential 1969 book co-authored by MIT’s Marvin Minsky suggested that they could not be very powerful.
However, not everyone is convinced by the skeptics, and some researchers have maintained the technique for decades. They were proven in 2012, when a series of experiments showed that neural networks powered by piles of big data can give machines new cognitive abilities. Processing so much data is difficult using a traditional computer chip, but switching to graphics card leads to an explosion in processing power.
