Search engines will usually be a part of most people’s experience when they sit down to use the Internet. But people tend to use them without really giving much thought to the nature of those systems. But a recent article published in the MIT Technology Review might make people rethink everything they thought they knew about the technology.
The article begins by showcasing how little search engines have changed in the past twenty years. An entire generation has been born and come of age during that period. And in that time search engines have almost always been a simple box that people enter text into. This text is then matched against a list of items within a database. For such a vital part of most people’s computing experience it’s remained fairly primitive. There’s a good reason for that though. The next step up from there is to use a more intelligent system which would make use of deep learning and other aspects of artificial intelligence. And even the most advanced computer systems of the modern era are only just now reaching a point where that’s viable. And when it works it can be amazing. It can perfectly replicate the feeling of window shopping. For example, when on a website one might like a pair of boots and be able to get suggestions about outfits which would compliment them. All based on how things would look together.
But again, one of the biggest issue is the sheer computing power needed to run visual search or make use of the associated artificial intelligence techniques. Many companies are working on answers to the issue. But one company, Slyce, has come up with a solution that’s currently running on a number of different platforms. Slyce recognized that people would want to run visual searches from very underpowered devices such as phones. But even the most powerful smartphones can’t compete with the power of a desktop system. And desktops themselves aren’t able to really run efficient deep learning systems. But Slyce came up with a way to use a combination of distributed computing and advanced APIs to enable even the most underpowered devices to use visual searches.
Slyce developed a seamless technique which allows any program to be easily modified to communicate with their advanced data servers. These advanced servers can run all of the calculations on images in order to process them for a visual search. The end results are then seamlessly integrated into the program which made the initial request. This essentially allows Slyce to enable even the most underpowered device to run programs which make use of visual searches and deep learning technologies.