Create 6 hours worth of content in 60 seconds… what’s the catch?

What is AI and why does it matter for publishers?

Broadly speaking, artificial intelligence simulating intelligent behaviour in computers; allowing it to perform tasks that were previously only achievable by humans, such as learning, interpreting data and, in the case of Snap Tech, seeing. Computer Vision is a branch of artificial intelligence, which focuses on teaching a computer how to “see” and interpret the world around it.

In this case, Snap Tech’s technology lets a computer look at an editorial photo, from something taken on the red carpet to a street style shot, and find similar looks for consumers to buy at home. But why focus on similarity without finding just the real thing? We know that that most of the dresses being worn by people in magazines are custom couture, super pricey, or sold out, so we focus on finding looks for in-stock similar looks from the high-street; focusing on cut, shape and fabric patterns to let readers get the look on the budget that works for them.

With over a decade of engineering behind this technology, as you can imagine this is top secret stuff. But without giving away too many secrets, they use a blend of mathematical heuristics and machine learning to interpret the image. Mathematical heuristics is defining a set of hand-crafted rules around what you’re looking at — from “what makes a dress a dress” to “what’s the difference between shorts and skirts?” This technology is blended with results from machine learning.

Machine learning works on the premise of “training data” — imagine someone showing you over 1000 images of different types of dresses, then showed you an image you’d never seen before containing a dress. You’d be able to point at it and say “that’s a dress.” That’s exactly how machine learning works, and the fun is dealing with edge-cases like “is a really long t-shirt a dress or a t-shirt?” and then getting into the details, like is this Breton or floral? The “artificial brain” that we’re training is called a Neural Network, and it’s a delicate balance knowing how much information to provide and how detailed to go, as if you get too prescriptive it’s known as “overfitting.” This is what makes artificial intelligence and computer vision so fascinating — there’s some serendipity in search results due to a computer’s interpretation of the image, and it’s all about getting the balance right.

Why is AI useful?

With over 23 million items of clothing analysed since they started, Snap Tech have saved editors hundreds of hours, and let them deliver solutions which would have been physically impossible in the past. That’s the beauty of artificial intelligence — it extends the art of the possible.

One of the projects that we delivered this year was for the BAFTAs. On the night, as your favourite style icons are walking down the red carpet, you could shop the look live. This wouldn’t have been possible in the past — it would have taken dozens of editors multiple hours to create get the look content. Now it’s all possible in just a few clicks. Snap Tech are proud to have created the world’s first live shoppable red carpet experience powered by artificial intelligence, giving consumers the chance to get the red carpet look for themselves in a way that’s never been possible before.

You can find out more about our solutions for publishers here, and if you want to know what’s possible for your particular publication, you can email us through our Contact Us page.