How Far Can Google’s Linguistic Analysis Go?

Google has introduced a number of advanced, impressive features over the years. When it was first introduced in 1999, it managed the then-unthinkable task of conveniently organizing the Internet, which we now all take for granted. It introduced advanced quality checks to rank content from different sites against each other, mapped out the world, revolutionized email, and helmed a million other breakthroughs I have no room to mention. But one of the most sophisticated advancements Google introduced is barely noticeable—its linguistic analysis algorithms—and it keeps growing more complex.

Language analysis is difficult for computers because it can’t be reduced to a numerical process. Algorithms function as a series of individual “either/or” processes, so it and most other computing functions can easily be reduced to numbers. With human brains, language is relatively easy to understand; we’re naturally wired to deal with it, and we’ve been immersed in it since birth. But to computers, words with multiple meanings, sentences with complex syntax, and illogical grammatical rules are all foreign, irreducible concepts.

Still, the engineering masterminds at Google are spearheading an iteratively introduced system to improve machine understanding of natural language on two fronts: one to decipher spoken human speech and one to analyze the intent behind search queries to find the best possible results. The degree of sophistication these breakthroughs are at is already astonishing, and the heights to which they’re projected to grow are almost terrifying.

The Rise of Semantic Search

One of the first major linguistic analysis breakthroughs from Google came in 2013 with the “Hummingbird Update.” This update was an algorithm change that introduced the ability to “understand” a user’s query. Rather than picking out keywords in a query and scouting the web for iterations of those keywords, the Hummingbird update allowed Google to look at the semantics of the query, analyze the intent, and then find appropriate results from there.

To someone searching for “great burger restaurants in Denver,” this change didn’t affect much. But it represents the beginning of Google’s explosive rise in sophistication.

Personal Digital Assistants

Google Now emerged as Google’s answer to Siri, the personal digital assistant in most Apple devices. As moderate rivals, these two assistant programs initially started as gimmicks. They could pick out most of the words you spoke, assign them to various programs, and speak back to you if you needed feedback like information or an answer to your question.

Today, these assistants are streamlined artificial intelligence programs, and if you haven’t tried using them recently, you should. Both programs are able to recognize human speech at an impressive level, and deliver highly relevant results from both a local and a global perspective, with in-device and Internet-based search functionality. This is important to note because these programs are small-scale artificial intelligence systems, and they’re designed to mimic certain processes of the human brain.

Robot Writers

When it comes to natural language analysis, Google is all about interpretation of human speech. Other companies, like Quill, are on the forefront of exhibiting human speech. By now, chances are you’ve read at least one journalistic article that was written by an artificial intelligence program that aggregated raw information from the web and wrote about it in a way humans can easily understand. Considering Google’s love affair with sophisticated artificial intelligence and semantic understanding, it’s only a matter of time before the company acquires one of these “robot writer” companies, or develops an algorithm-based writing program of its own.

Ray Kurzweil’s Futuristic Visions

All these advancements are impressive, and help you understand why every Google search you perform seems to find the perfect set of results, but what’s really interesting to SEO experts and digital marketers is how far this technology has left to grow. For that, I look to Google Head of Engineering Ray Kurzweil, whose futuristic predictions are straight out of a dystopian science fiction novel. In one of his most popular TEDTalks, Kurzweil discusses how artificial intelligence is being created with low-level processes called modules, which together form higher-level conclusions and analyses capable of semantic understanding. He predicts a world when soon, search engines will make predictive suggestions and announcements to you based on what it knows about your personality and interests. He predicts a world not long after that where human and artificial intelligences become indistinguishable—essentially, you’ll be able to “think” your own Google searches.

According to Google’s lead algorithm engineer, the answer to the title question—how far can Google’s linguistic analysis go—is simple. It can become more sophisticated than humans, to a point where it hybridizes with humans. At some point, the Internet will seek to be some other place, and will simply be a heightened form of reality, and searches won’t matter because we’ll be plugged directly into the cloud. This would definitely put a damper on your SEO campaign.

One Year at a Time

Despite being grounded in reality by technologists and scientists, this prediction still seems a little far-fetched. If, in 30 years, Google’s algorithm advances supersede humanity as the dominant thinking force on the planet, it would certainly put the kibosh on online marketing (not to mention, most industries as we know them). But the future is far off and unpredictable, so all we can focus on is the now.

Rather than speculating about what Google could be capable of, look at what it is capable of, and take things one year at a time. For now, Google rankings still matter—a lot—and you can get yours, in part, by writing great content, optimizing for your geographic location, and ensuring your site is optimized for mobile. Next year may bring something different in terms of semantic search capabilities, but that’s for next year to decide.




Source: How Far Can Google’s Linguistic Analysis Go?

Via: Google Alerts for AI