The Learning Machine
For much of human history, machines were simply used as a tool for humans to do things. Without us directly controlling them or monitoring them extremely closely, they wouldn’t work or could cause major problems if something goes wrong. However, soon, advancements led to humans being able to take a more “hands-off” approach when using technology, as some machines could work on their own, like the kind you find in factories all over the world.
Now, machines like that are the standard and are found in factories all over the world. Also, creations like the phonograph, photograph, steam engine and Spinning Jenny were able to do things that people could not have even fathomed when they debuted.
Just like they were back then, modern machines are now capable of things we never thought possible, even a few years or decades ago. Here are a few examples of things machines can do that (or are learning to do) that we never thought possible before.
Clean our Homes
This isn’t really cutting edge or anything, but I included to show just how quickly something can go from being “amazing and new” to “normal part of life”. When the Roomba first came out in 2002, nothing like this had been available in a mass market before. This little robot would intelligently crawl throughout your home and clean the floors of any debris without you having to do a thing. But a little over a decade later, and having sold over 10 million units worldwide, they are just a normal part of life, which shows how quickly amazing technology and machines can be accepted as the norm.
Learn From Humans
We have been able to program computers and machines for quite a while to make them “learn” new things. However, that is not real learning. But now, machines are truly becoming “self-programming” through deep learning. For example, some machines can be given an input and a known output and the machine will itself learn the optimal algorithm. This will allow it to predict the same output with new inputs, on its own.
However, even with all this progress, we are still a ways away from machines matching human learning patterns and truly knowing how the machines learned what it did. Even still, no one a few decades ago would have thought machines could learn new things without being programmed.
There you have it, some machines that can do things that are truly extraordinary. Just like the machines of the industrial revolution are now the standard, look for these technologies/machines to eventually be something we don’t even bat and eye to. And by then, even more autonomous and amazing machines will have been created that can do things we would never expect now.
Analyze Shopping/Fashion Trends
IBM has been on the cutting edge of predicting fashion and shopping. In terms of fashion, IBM’s Watson has analyzed hundreds of thousands of designs and aggregates factors such as color, fabric and hem length to predict trends. They even had some clothing appear at Melbourne Fashion Week.
In terms of shopping, the IBM acquisition of The Weather Channel allowed them to create Deep Thunder. Deep Thunder is a customized forecasting model that looks to predict consumer shopping patterns and help companies adjust accordingly to these patterns.
While these technologies/machines are likely years and years away from coming to market or even being used outside a lab, the fact that they are even possible is mind-blowing to some people. A process called entanglement is making some think that teleportation isn’t impossible. As far as invisibility goes, a cloaking device using metamaterials has successfully been able to make microscopic objects become invisible.
By Kale Havervold