Of all technologies developed in the past decade, nothing has been quite as influential as AI. What was once a topic of science fiction novels and movies has become all-too-common. So much so, that AI has been begun managing our daily lives before any of us started noticing. Everything from our TV, our smartphone's face unlock feature, selfie filters, our navigation apps are all powered by artificial intelligence.
On a grander scale, artificial intelligence also powers what companies serve to us. Google uses artificial intelligence to serve ads better, increasing revenues from online shopping to unprecedented heights. Facebook, who extensively uses AI to recognize people in the photos, will also be using AI to generate fashion recommendations. AI has revolutionized not only how the largest tech companies work, but the companies themselves, and this is all thanks to the rapid developments in AI the past year.
What is AI?
But first, the question is, what is AI? While AI is a broad term that has been thrown about in fields such as video games, Artificial Intelligence, or AI are usually algorithms that are used to simulate intelligence. AI has been developing for the past few decades, but it was developments in a certain type of AI that started a revolution in the world this past decade: Deep Learning.
Deep Learning is a type of machine learning. While Machine learning is a process in which a computer is aimed to organically teach itself, often by trial and error, to solve a problem without being taught a solution, Deep Learning, on the other hand, is layered machine learning. With an algorithm on one side that accepts input, one or more hidden layers, and an algorithm for output.
Hidden layers are where the magic happens, and they're what many would call a "black box". Machine learning algorithms use methods like pattern recognition, obtaining insight from large amounts of input, and start making guesses as using these insights. If these insights are inaccurate, they are trained again to find new insights or are tuned with different parameters until they start guessing accurately. Eventually, these guesses become accurate enough to detect small details such as health warning signs that doctors might miss and suggest to them a possible diagnosis for Acute Kidney Injury.
Modern Deep Learning
Deep Learning and AI again aren't new, in fact, they were all the rage among computer scientists as early as the 1980s. And discovers during this time laid the groundwork for developing the modern AI, but it wasn't until this decade that AI was actually developed rapidly.
In 2010, Fei-Fei Li's ImageNet launched the Large Scale Visual Recognition Challenge (ILSVRC), providing a large dataset for image recognition for training and validating AI results. The competition brought AI to the spotlight, and everything began snowballing.
Every major tech company in the world began taking its first steps to develop Artificial intelligence. In 2010, Apple acquired Siri, so today, their smartphone assistant powers every iPhone. Ian Goodfellow Apple's director of Machine Learning, although he was still with Google at this point, also developed Generative Adversarial Networks (GANs), where artificial neural networks competed against itself in order to improve.
Today, GANs are used to create DeepFakes, frighteningly believable AI-doctored videos where people, often public figures, are imposed unto other people and made to act or speak things that they never did. Worryingly the number of DeekFake videos rose as much as 84% in only seven months.
Of course, not everything is doom and gloom with Artificial Intelligence. Thanks to amazing developments in DeepLearning, Google's Waze now help drivers comfortably and safely get to their destinations faster. AI is used by from Boston University School of Public Health to spot potentially unsafe food products on Amazon. And AI will not only be powering self-driving cars in the future, but Stanford researchers are also already teaching AI to drive like a race car driver and adapt quickly to extreme conditions.
Good or bad, AI is, without a doubt, the single largest innovation in the past decade, and 2020 will see a new era where AI will truly be part of every aspect of human life. Now, can I have my own Jarvis yet?
