Deep Learning, In Brief
We can’t talk about the impact of deep learning on search and online marketers without talking about what, in fact, deep learning is. I wrote a pretty in-depth explanation in this post: Why Deep Learning (and AI) Will Change Everything but I’ll give you the cliff notes version:
- Deep learning utilizes clusters of virtual neurons to extract insights and statistics from large data sets. Deep learning creates layers of processing steps that transform data in a nonlinear fashion to make predictions, and is modeled after the human brain. What’s really cool is that the more data these ‘smart’ clusters analyze, the more accurate their predictions.
- Deep learning falls under the AI umbrella and is closely related to machine learning. Machine learning, however, relies on humans to input commands or analyze the output. Deep learning, on the other hand, is unsupervised.
- Thus far, deep learning has mainly been used in speech recognition, language, and vision processing in both computers and robots. Many startups from varying industries have exploded onto the deep learning scene and new applications are popping up with space-race-esque fervor.
Deep learning is truly a technology still in its infancy, but so much is on the horizon here, it’s really fascinating.
Now, onto the main event: What’s all this got to do with SEO?
Deep Learning and Search: What You Need to Know
Deep learning is poised to revolutionize search. Facebook and Microsoft (along with Chinese search engine Baidu) have all thrown their hats in the ring on this front, and that speaks volumes. Facebook has devoted an entire business segment to AI FAIR— (Facebook AI Research)—and boasts facial recognition software, programs that can recognize and categorize actions in videos and much more. Microsoft has Project Oxford, a cognitive services segment with the tagline “Give your apps a human side” that caters to those hoping to integrate vision, speech, language, or search APIs into their existing platforms.
Oh, and there’s Google. Google recently added deep learning-based RankBrain to its staple Hummingbird algorithm, changing search as we know it. (Read more about RankBrain and how it affects SERP placement in this piece: The RankBrain Impact on Search: Why You Should Care)
Those are some of the main players—let’s break down the game. Below are a few things to watch for as deep learning storms the search scene:
More SEO verticals, more opportunities. Deep learning gives marketers the opportunity to leverage new channels that specialize in more than just text. By leverage, I’m talking about getting SEO love for snippets of non-text media that fulfill not only searchers’ queries but also acknowledge their intent. THAT is the key. With deep learning, any platform that is rich in visuals (images, presentations, videos) is fair game as an SEO vertical, presenting mega opportunities previously untapped by online marketers.
More AI devices, more search integration. You can fully expect the impending AR/VR device deluge to require smarter search engines—ones that, in a sense, understand what sites pair best with which devices and use perceptive ranking technologies to deliver. The implications here are far-reaching.
Fewer spam sites, less ‘white noise’ content. Sites that use spam tactics won’t clutter SERPs much longer thanks to deep learning. Remember, deep learning does more than recognize patterns in data—it draws insights along the way, getting smarter as it goes. So, all those web pages full of redirects or white noise content—content that serves little value except to exist on a page—will be the first to drop off deep learning’s radar when it comes to ranking.
No more anonymity. This may sound a bit Big Brother-ish, but AI paired with search will make it increasingly difficult—not entirely impossible, but difficult—to retain any sort of substantial online anonymity. To that end, new supercookies have already been released by Verizon, AT&T, and Facebook, bringing what were once futuristic predictions front-and-center today.
It’s quite clear that if you’re a marketer or SEO pro and you haven’t yet interfaced with deep learning, you’re on course to do so sooner rather than later. How will your business stay ahead of the curve? Is there anything my colleagues and I could cover that would help along the way? If so, let me know in the comments. I’d love to hear your thoughts on deep learning and any experiences you’ve had thus far with this up and coming technology.