26 Jul 2017

WRITTEN BY: NATHAN SIKES
TechCrunch posted this today. We are huge fans of TechCrunch and artificial intelligence research

The concept of deep learning or deep structured learning has been a frequent topic of conversation in recent months because of the commitment and advancements of some of the world’s largest and most prolific search companies. With organizations like Google, Facebook, Microsoft and Baidu (a Chinese search engine) buying into this technology, we are starting to see a huge acceleration of the applications and uses for this relatively new artificial intelligence (AI).

The focus of this article targets the technology of deep learning and its influence on search engine optimization (SEO) in today’s online world. Let’s take a look at what deep learning is, some history and some of the major players who are using this AI to shape and improve the digital landscape around us.

The Quest For Real Artificial Intelligence

Before we can truly grasp the implications of deep learning, we need to better understand what it is and where it came from. Although you can’t track its origins to one particular person or place, most authorities in the field agree that Geoffrey Hinton is the godfather of modern-day deep learning. Currently splitting his time between Google and the University of Toronto, Hinton’s contribution to Backpropagation in the 1980s, and most recently the Neural Computation and Adaptive Perception Program, has re-ignited a flat industry in need of a spark. “Over the last 20 to 30 years, he [Hinton] has been pushing forward the frontier of neural networks and deep learning,” says Kai Yu, the director of Baidu’s Institute of Deep Learning. “We have never seen machine learning or artificial intelligence technologies so quickly make an impact in industry. It’s very impressive.”

Many companies are seeing that the future of deep learning is here, and that it doesn’t require a lot of money or resources to take advantage of this new industrial science.

Hinton’s contributions can be tracked to almost every large entity currently working on artificial neural networks and deep learning. Yann LeCun, currently at Facebook, helped develop Backpropagation alongside Hinton in the 1980s. Andrew Ng, the Chief Scientist at Baidu, founded the Deep Learning Project at Google in addition to working there with Hinton for years. Although these handful of engineers may be working for competing companies, they’re all desperately in search of the same thing — real behavioral learning in artificial intelligence.

Previous to the last decade, the business industry had all but turned its back on the field of deep learning. The limitations of chip-processing abilities and the data sets used in the artificial neural networks that ran them made Hinton’s and his colleagues’ theories impractical and ahead of their time.

Fast forward to 2015 and we see a completely new environment full of potential applications, and the big boys aren’t the only ones who noticed. Smaller organizations like Clarifai are using augmented intelligence for advertising, moderation/filtering and content curation purposes. Other companies like Moz (an SEO company) also see this as the future, and a way to provide better products and services to their customers.  “A lot of technology firms like Moz do some level of machine learning. We don’t do anything with deep learning and a lot of neural networks. We might be going that direction,” Rand Fishkin explains on the Moz blog.

Many companies are seeing that the future of deep learning is here, and that it doesn’t require a lot of money or resources to take advantage of this new industrial science. IBM’s Watson Analytics offers a freemium service that allows you to upload up to 500MB, and enables you to explore your own real-life applications for deep learning. Inputting Google Adwords or other sales metrics into this tool can help even startup companies find relational and predictive information in their data. In addition to Watson Analytics, let’s take a look at the other companies developing and using these technologies right now.

Google Research: Google ranks third on Forbes’ list of “The World’s Most Valuable Brands In 2015,” but I don’t think many would argue that they’re first in search. Google has made many advancements over the past decade when it comes to machine intelligence, and has been working on building a better understanding of images, videos and language.

Through their research, acquisitions (think DeepMind, acquired in 2014) and partnerships like the one with Imagenet Large Scale Visual Recognition Challenge, Google has remained committed to testing and applying new areas of deep learning. Most recently, Google announced they now have the ability to automatically produce captions and describe images the very first time they see them. Can you imagine if image recognition or features like this were implemented within Google’s search algorithm?

It may not be that far off. Geoffrey Hinton explains in a Reddit Ask Me Anything session, “I think that the most exciting areas over the next five years will be really understanding videos and text. I will be disappointed if in five years time if we do not have something that can watch a YouTube video and tell a story about what happened.”

Facebook FAIR: The most popular social networking site in the world has become a strong player in search, and will continue to grow their market share in the years to come. Facebook AI Research (FAIR) shows Facebook’s commitment to artificial intelligence and deep learning as the future of social, purchasing and media. In fact, one could argue that between their facial recognition software, their contribution to open source modules for Torch (a development environment for deep learning) and the most recent announcement by Mike Schroepfer (their CIO) stating that Facebook’s AI now has the ability to recognize actions in video, Facebook is the current leader in utilizing deep learning intelligence.

Microsoft Project Oxford: Microsoft recently made waves with their how-old.net facial recognition project. What most people don’t realize, however, is that this undertaking was designed by Microsoft’s machine learning research team. In fact, according to Microsoft’s blog, “It took a couple of developers just a day to put this whole solution together, from the web page to the Machine Learning APIs to the real time streaming analytics.” And that’s only the beginning. In addition to projects like this, we will also surely see an integration of Project Oxford and Cortana (Microsoft’s “personal assistant”) into Windows 10 and Edge, the Internet Explorer replacement.

WolframAlpha: WolframAlpha isn’t a household name — yet — but has been a big player in the artificial intelligence industry for the last several years. Their ultimate goal is to compute anything and everything, but they focus primarily on fields that require expert-level knowledge or capabilities. Recently they’ve expanded their reach to image identification, problem generators, language recognition and even analytics for Facebook.

So Why Should I Care?

We have finally passed the threshold of if artificial intelligence will be integrated into the search to when it will be, and the companies mentioned above are leading the charge. So given this new information, what kind of impact will deep learning really have on search over the next five years? Here are a few predictions:

New SEO verticals: If you’re an online marketer and this technology doesn’t get you excited…you may be in the wrong field. This type of artificial intelligence offers search engine marketers a chance to take advantage of a variety of new channels, including image-heavy social networking sites, video-sharing sites and even slide presentation platforms, while getting “credit” for images and videos that bring value to searchers.

It’s becoming increasingly more important to capture your customer’s imagination and attention with visuals, and search companies are taking notice.

Death of the spam site: Google has been working on fighting spam sites for a long time. They even have an algorithm called Penguin that is dedicated to decreasing the rankings of sites that use spam tactics. My prediction is that we will see a huge decrease in websites that use spammy techniques, sneaky redirects or have thin content with no value. This will provide a safer, less cluttered web.

Better device integration: Google’s mobilegeddon update wasn’t the last of its kind. As companies like Facebook (Oculus) and Microsoft (Hololens) put out their virtual reality headsets, we are going to need a smarter search engine that has the ability to prioritize and learn what websites should be displaying on which devices.

No place to hide: As search engine companies perfect this AI and become more intelligent, so will their tracking capabilities. We are going to see a new generation of “supercookies” like the ones Verizon, AT&T and Facebook have recently released, which is going to make it increasingly more difficult to stay hidden online.

Visual content will be more important than ever: Millennials ignore ads on a greater level than any other generation before them. According to Simply Measured (a social media analytics company), 62 percent of all brand posts and 77 percent of all produced engagements on Facebook are photos, as opposed to just normal text. To go even deeper, Hubspot’s research says that you can generate up to 94 percent more views and 37 percent more engagement if you add compelling visual elements and graphics to blog posts and social media content.

It is becoming increasingly more important to capture your customer’s imagination and attention with visuals, and search companies are taking notice. Look for this type of content to be prioritized in the future.