Artificial intelligence tech is everywhere these days, informing everything from credit decisions to diagnosing diseases and keeping cars from crashing. TED-Ed’s Briana Brownell explains the three main kinds of machine learning technology, how they differ, and how little we really know about how AI works.
Awesome Artificial Intelligence
After multiple lockdowns in the UK, vlogger Tom Scott is starting to run out of ideas of things he can make videos about. So he turned to an artificial intelligence to see if it could come up with any. OpenAI’s GPT-3 tech came up with a mix of mundane, ridiculous, and surprisingly legitimate-sounding ideas, depending on its tuning.
What happens when you let artificial intelligence write a Christmas song? Musician Chase Holfelder wanted to find out, so he turned to the Jukebox neural network to create the lyrics, chord progression, and vocal melody of this tune, titled “Rudolph, the All-Gracious King.”
Developed by Magenta using Google AI tech, Tone/Transfer takes ordinary sounds like a human voice and makes them sound like a musical instrument. It can also digest the sounds made by one kind of musical instrument and map them onto a different one. You can play with the online demo for yourself.
Computer scientists from Google, MIT, and UC Berkeley have developed incredible image correction software that uses artificial intelligence to remove or soften unwanted shadows and harsh highlights from faces without affecting other naturally-occurring shadows. Read more about the tech on the project’s website.
It’s the year 2028, and while the world goes about its business, a system basically eradicates all knowledge of 20th century pop culture. Tom Scott explores one of the negative possibilities of artificial intelligence run amok in his own 6-minute episode of Black Mirror.
Machine learning tech is getting pretty good at things like object recognition, but can it write a decent song? Funk Turkey fed artificial intelligence software a bunch of AC/DC lyrics and asked it to create a new track. This is what it came up with. We’re pretty sure this AI is actually a 10-year-old boy.
Artificial intelligence is getting better at identifying objects in still images, and more recently in video. Now machine learning tech is getting smart enough to look at what’s happening in a video and answer questions about what it has seen. Two Minute Papers provides a brief overview of CLEVRER and its capabilities.
Looking to create your next big hit song? This AI-powered song lyric generator will craft the words for you automatically. Simply select a topic, choose a genre and a mood, and it’ll take care of the rest. Sure, your masterpiece might not make much sense, but that never stopped Duran Duran or David Bowie.
We’ve seen machines that can sort LEGO bricks before, but they’re generally limited to just a few specific shapes or colors. Daniel West’s machine is much smarter, using AI algorithms to identify and sort nearly 3,000 different LEGO shapes and colors. We think it’ll need more than 18 sorting bins to be really useful though.
Season 8 of Game of Thrones had more than its share of haters and controversy. So in the interest of improving its ending, College Humor turned to artificial intelligence tech to rewrite bits of the script for the series finale. At times it’s nonsensical, but at others, it’s better than the ending we got.
Next top model? How about an army of free and tireless ones? A Japanese company called DataGrid claims it has created a way to automatically generate whole body models of people who don’t exist. These virtual models can be given a variety of poses and clothing.
Need to remove the background from an image of a person? You could try and do it yourself in Photoshop, or you could use this website, which uses AI tech to distinguish between people and backgrounds. It does an impressive job, though its output resolution is currently quite limited.
Artificial intelligence isn’t just powering robots, it’s already making decisions about our credit, work, insurance, safety, and other important aspects of our lives. Dress Code and HPE interviewed experts about some of the issues raised when algorithms wield such power.