Changes in the way search results are linked on Google are on the way, and this may make our search results more accurate.
Google is one of the most famous search engines on the internet, if not the largest, obviously there are very interesting solutions like DuckDuckGo, for example. Alis, we have a super detailed post about this alternative to Google. Check out, along with this subject, there are videos demonstrating various features and characteristics of the search engine.
Googlando on Google
Doing Google searches has become part of our daily lives, however, the verb googlar He started composing as part of our language. Even with millions of people using the search engine, the vast majority do not know how it works. Obviously, I'm not talking about technical terms or the programming used in your algorithm. However, knowing how to search (as little as possible) can speed up and even enhance your results on the internet. A very interesting video, which is indicative of the channel Bson Training. Not to be overwhelming in demonstrating the various commands I know about Google, I remembered your video on the subject. If you want better results, consider watching the video and learning 15 commands to use Google search more efficiently.
Google's search engine currently uses natural language processing to parse queries. This means that if anyone searches for: get medicine at the pharmacy, Google will display results for words medicine and pharmacy. All web pages indexed with these keywords will be displayed, and this is not always the most accurate. After all, the context itself is not considered, just loose words. Another point, are words that refer to others or have another meaning. For example, in the case of the remedy Google could also suggest web page results with the word drug.
Starting today, Google becomes smarter by using a new language processing method. Through machine learning, the new feature promises to make the results as close as expected. Called from BERT (Bidirectional Encoder Representations From Transformers, as the Transformers Bidirectional Encoder Representations) is the solution capable of parsing word sequences and not just each word in isolation.
To get a sense of BERT results when searching for can i buy some medicine at the pharmacy The result considered the context and indicated an article from the Department of Health and Human Services demonstrating how to request medication from a friend or family members. In the traditional way, the result was not so assertive and led the user to issues related to how to obtain a filled medical prescription and eventually missed the point of the research.
At a press event held yesterday, Pandu Nayak, Google’s vice president of search gave another example, with the following search: How old was Taylor Swift when Kanye took the stage?. Prior to BERT, Google offered several results, being videos from the 2009 event, pop star acceptance speech at MTV Video Music Adwards. After Bert, the result was straight to the starting point of the question, Google first ranked in the ranking an excerpt from a BBC article, highlighting the age of the singer considering the context of the issue.
At the front of the ranking, this is the biggest change we've had in the last five years and one of the biggest since the beginning., said Nayak.
When questioned about the effectiveness of such improvements, Nayak says BERT would have real value in more exotic questions and would affect only one in ten surveys in the United States. Still, with a monstrous daily search traffic of its more than 90.8% in the search market across all its products, hundreds of millions of results would change overnight.
As we answer more exotic questions, I hope this will lead people to ask more and more exotic questions., he said.
However BERT is not perfect, and there are areas that it does not handle well. When searching for Tart (English tartan), a checkered pattern of prints, the result was not so pleasant and images were ignored. Without the use of technology, such figures representing the fabric were highlighted for the end result.
Another case was in the search Which state is south of Nebraska?, BERT presented the Wikipedia page to South Nebraska (a neighborhood of Florida), by contrast, the ideal result would be the Kansas Wikipedia page.
Initially only available for English and US research, BERT will be gradually incorporated into all research in the country and in the future there are plans for other countries and languages, Nayak is confident with this technology.
We are playing a statistical game here. Together we know that any change will have some gains and some losses. I think the (BERT) reward is very good.
Change can attract more traffic within the search engine and impact company services. Google Maps, Youtube, among others may highlight niche searches and therefore content creators can benefit, but all uncertain and only in the coming months we can see the impact caused by BERT, first in the US and then, who knows. , in other countries.
What's your take on Google's search engine and this new onslaught?
Until the next post, I wait for you here on the blog Diolinux, SYSTEMATICALLY!
See an error or would you like to add any suggestions to this article? Collaborate, click here.