Machine Learning: Artificial Intelligence
The ability to solve problems and to adapt to surroundings is generally called intelligence. This ability is naturally restricted to biological life forms. In science fiction, intelligent machines such as Data from Star Trek: The Next Generation were thought of. New developments in the computing world could lead to computers that adapt to surroundings and solve problems, like in Star Trek. Therefore, intelligence would no longer be restricted to naturally occurring life. The still-developing technology of machine learning and artificial intelligence has many theoretical applications and is very likely to match human intelligence in the near and distant future.
Artificial intelligence (AI) is the science of making computers perform tasks that would normally require human intelligence to perform. Human intelligence consists of "learning, reasoning, problem-solving, perception, and language-understanding" (Copeland). Intelligence of this kind is usually only found in nature, like in humans and a few other animals such as monkeys.
Google Brain was a network of 1,000 computers that could sort images into different categories. After three days of searching through photos, it concluded that the internet is mainly full of humans and cats (Jones). Although computers are fast enough to go through 10 million photos in three days, it took many computers and plenty of code to identify subjects of images, a task that one human could do with no aid. Because of this, Google Brain is considered a major achievement in the category of machine learning and AI. Google Brain's many computers mimicked the neurons of the human brain. This is called deep learning. Google Brain's deep learning network was ten times larger than any other network before it (Jones).
Another neural network was trained to find fossils based on satellite images. Bob Anenome, the paleontologist training the computer network to find possible fossil sites, gave it satellite images of terrain, and marked whether a pixel of the image was a fossil site or not. The computer compared the aspects fossil sites and non-fossil sites in six different light wavelengths to deduce which pixels of an image were likely future fossil sites or not. When exposed to unfamiliar satellite images, the computer identified 79% of the pixels known to be fossil sites. Of these pixels, 99% had fossils. This program could be used with any terrain, as long as the computer has been trained on an image of similar terrain (Callaway).
Machine learning may also be used to process big data. In the human mind, many tasks are performed simultaneously in different regions of the brain. The same applies to neural networks. The data could be split up for many computers to each do their own part (Quellette). This would be very efficient technology.
However, such technology, in thinking like humans, is feared by a prominent scientist. Famous physicist Stephen Hawking stated that AI would "take off on its own, and re-design itself at an ever increasing rate" (Cellan-Jones). Elon Musk, founder of Tesla Motors and SpaceX, also shares these views. He says that AI could be "more dangerous that nukes" (Sofge). While there is a possibility of smart robots turning against us and obliterating humankind, the technology necessary for such a rebellion is in the far future.
Intelligent machines could aid us greatly. AI could save our lives. It could even match humans one day. While today's AI is not intelligent enough to match humans, future AI could very well match or surpass humanity.
Works Cited
Callaway, Ewen. "Artificial Intelligence Finds Fossil Sites." Nature.com. Nature Publishing Group, 8 Nov. 2011. Web. 21 Mar. 2015.
Cellan-Jones, Rory. "Hawking: AI Could End Human Race." BBC News. 2 Dec. 2014. Web. 21 Mar. 2015.
Copeland, Jack. "What Is Artificial Intelligence?" AlanTuring.net. May. 2000. Web. 21 Mar. 2015.
Jones, Nicola. "Computer Science: The Learning Machines." Nature.com. Nature Publishing Group, 8 Jan. 2014. Web. 20 Mar. 2015.
Quellette, Jennifer. "How Quantum Computers and Machine Learning Will Revolutionize Big Data | WIRED." Wired.com. Conde Nast Digital, 14 Oct. 2013. Web. 21 Mar. 2015.
Sofge, Erik. "Why Artificial Intelligence Will Not Obliterate Humanity." Popular Science. 19 Mar. 2015. Web. 20 Mar. 2015.