Google’s BERT Update: New Google AI For Better Search Results | Bloggercage.com

Google Bert Algorithm UpdateGoogle Bert Algorithm Update

Find out how BERT, Google’s new Artificial Intelligence system, will affect the SERP’s and the results of your online strategy!

Google will make the biggest change to its search engine since the introduction of RankBrain – almost five years ago. The company announced that the action will impact 1 out of 10 pages of results, in order to substantially transform the results that match those searches.

The so-called BERT, named after Bidirectional Encoder Representations from Transformers, was launched this week for English language searches and will soon be available worldwide. 

Want to know more about what’s new and how it will impact search? So stay with us and check it out!

What is BERT and how does it work?

BERT is a neural network technology that aims to improve natural language processing, as commonly used by users in search engines. That is, it helps computers understand the language more similar to human understanding.

The program is used to grasp the context nuances of the words entered in the search and to find the most appropriate and relevant results.

Until last year, the program followed the open-source model and has a detailed description of its operation on the Google Artificial Intelligence blog.

To illustrate how the solution works, in a search such as “2019 brazil traveller to the USA need a visa” the word “to” and its relation to the other search words are important for understanding the meaning of doubt.

Previously, Google could not understand the importance of this connection and would return with results about US citizens interested in traveling to Brazil.

“With BERT, the system is able to grasp this nuance and know that the word ‘to’, which is quite common, is actually essential in this context and so it is possible to suggest much more relevant results,” according to with the company.

Note: The examples used are for illustration purposes and may not correspond to the reality of real-time searches.

wyX1R32gQhPgqhxC 7anO3g8W7uDxDDBLM9dbQILoG8xCSYPEgt3RDDY24Y3fgiDQlcqcdoifTJfVZmabStmXmEcxGFOhNL3GiDSbcCK ByBrAgMb8cUz6ssfwVe0gkUHEXCZoWTwyX1R32gQhPgqhxC 7anO3g8W7uDxDDBLM9dbQILoG8xCSYPEgt3RDDY24Y3fgiDQlcqcdoifTJfVZmabStmXmEcxGFOhNL3GiDSbcCK ByBrAgMb8cUz6ssfwVe0gkUHEXCZoWT

In another example, in a search for “Can you get medicine for someone pharmacy,” you can see much more accurate results according to the user’s intent.

wyX1R32gQhPgqhxC 7anO3g8W7uDxDDBLM9dbQILoG8xCSYPEgt3RDDY24Y3fgiDQlcqcdoifTJfVZmabStmXmEcxGFOhNL3GiDSbcCK ByBrAgMb8cUz6ssfwVe0gkUHEXCZoWTwyX1R32gQhPgqhxC 7anO3g8W7uDxDDBLM9dbQILoG8xCSYPEgt3RDDY24Y3fgiDQlcqcdoifTJfVZmabStmXmEcxGFOhNL3GiDSbcCK ByBrAgMb8cUz6ssfwVe0gkUHEXCZoWT

The program will also impact Featured Snippets – which will already have their searches changed in all languages. 

To illustrate this impact, here is an example where Google presents a much more relevant snippet to answer the question: “Parking on a hill with no curb”. In the past, such a search would cause real confusion in the search algorithm. 

According to Google, “We often put a lot of attention on the word“ curb ”and not the word“ no ”without understanding how critical this term was to enable an adequate response. So we came back with results about parking on a curb hill. ”

wyX1R32gQhPgqhxC 7anO3g8W7uDxDDBLM9dbQILoG8xCSYPEgt3RDDY24Y3fgiDQlcqcdoifTJfVZmabStmXmEcxGFOhNL3GiDSbcCK ByBrAgMb8cUz6ssfwVe0gkUHEXCZoWTwyX1R32gQhPgqhxC 7anO3g8W7uDxDDBLM9dbQILoG8xCSYPEgt3RDDY24Y3fgiDQlcqcdoifTJfVZmabStmXmEcxGFOhNL3GiDSbcCK ByBrAgMb8cUz6ssfwVe0gkUHEXCZoWT

BERT versus RankBrain?

Despite the innovation of this system, it is important to note that RankBrain is not dead. Remembering: RankBrain was Google’s first Artificial Intelligence-based method for understanding searches.

It relies on analyzing both search and content on Google-indexed web pages to better understand the true meaning of words within a given context.

Thus, BERT does not replace RankBrain but acts as an additional method for understanding content and search. When Google concludes that the question will be best answered with the help of BERT, it will be used. 

Also, it is important to remember that in a single search, you can use more than one method, including BERT, for a better understanding of user intent.

And how does it work? Google explained that there are a number of ways to understand what search language means and how it relates to web content. 

For example, if you type something wrong, error detection systems help you spell a word correctly so you can find what you need.

Similarly, if you use a synonym for the word you are looking for, it can find content that spans both terms.

Thus, BERT is just another sign that Google uses to understand the language. Depending on which search you perform, any of these systems, or their combination, may be used to provide the best available result.

Is it possible to optimize for BERT?

The answer to this – until then – is only one: it is unlikely. Google has stated that SEO experts cannot really optimize even for RankBrain. 

But anyway, that means Google is becoming better and better at understanding natural language. So keep producing quality content for your audience, not for the algorithm. 

It’s important to keep an eye on this update not only because Google itself considered it “the biggest leap in progress in the last five years, and one of the biggest in search engine history,” but also because 10% of all searches were impacted. by BERT.

With that in mind, we recommend that you keep an eye on changes in traffic as soon as the change takes effect for the Portuguese language as well, and share with us if there is any impact on your pages.

This content is a translation adapted from the text of Search Engine Land.