Computing has become a central part of most people’s lives. If one were to ask most people about their daily activities than it’d be certain that most would talk about getting online to search out information. It’s such a common part of people’s lives that most never stop to think about how little it’s changed over the years. Most aspects of the Internet have changed and evolved since the Internet first gained real popularity. But the process of going online and typing text into a search engine really hasn’t changed much since the 1980s. But a story in the MIT Technology Review suggests that might finally be improving.
The article begins by pointing out how little things have changed in search engines over the past few decades. It then goes on to discuss a new take on how people search online. And this is through a combination of advanced artificial intelligence called deep learning that’s combined with visual processing systems. This combination is known as visual search. It’s functions in a way that’s far more similar to how people search for information in the real world. When people look for similarities in objects it’s usually done by looking at things. And these new visual search tools are able to use images to replicate that process.
However, one of the reasons this has taken so long to start showing up in the world is due to the amount of processing power needed. It’s far more computationally intensive than what smartphones, desktops or even most companies webservers are capable of. While many companies are trying to address this issue, one in particular called Slyce has developed a drop in solution. They’re using a method of distributed computing to essentially allow people to use their advanced data servers as secondary processors thanks to modern high speed internet connections.
Slyce has been able to create a simple way to instantly turn any system into a high end visual search tool. One example of this is their universal scanner. This allows people to use a smartphone to take a picture of an item, such as a purse, and instantly search through a company’s catalog to find the same or similar purses. And this can be extended to any item within the inventory. And because of the distributed nature of the system it’s easy for Slyce to roll out upgrades to constantly add new features and performance tweaks to the system.