In 2015, Google integrated RankBrain into its search algorithm, an artificial intelligence that will further improve the results it has displayed and also display results that match its own for previously unknown searches. What RankBrain really is and how Google's artificial intelligence works, explains Alona Demchyk in the following guest article.
The fact is artificial intelligence (AI) technology is expanding into ever more increasingly. A driving force behind this development is the Google parent company Alphabet, which inscribes artificial intelligence ever deeper into the DNA of its all areas of life covering subsidiary companies. The company's product portfolio includes medical technology (contact lenses for diabetics), mobility (drone delivery, self-driving cars), renewable energies (wind turbines, so-called "energy kites") and smart home technology (NEST) (Source: All of Google's , Alphabet's-companies and products from A to Z ).
In the everyday life of most people, however, the Google search engine is still the most important point of contact with the company from Mountain View. Whether it's on the desktop, mobile phone or smart TV, Artificial Intelligence and Machine Learning have now made a significant contribution to choosing the right results on billions of searches every day.
This is particularly clear on a part of the Google algorithm, which officially carries the somewhat strange name "RankBrain".
The first official announcements for the use of artificial intelligence under the name RankBrain, there was Google in early 2015.
The system, it was said, help to interpret the search results on unknown, new searches and more complex inputs from the natural voice search correctly and appropriate To deliver search results.
At that time, according to Google, about 15% of all Google searches were affected. Meanwhile, it is active in the background with all around three trillion searches per year.
Machine learning, in its purest form also called "Deep Learning", is an automated learning process in which a system can generalize the learning data by pattern recognition, comparisons and examples and apply it to new, unknown situations.
hind RankBrain hides an artificial neural network, which learns without human instruction, from existing data and information assignments and relationships to recognize and use.
But such a system can only be called Artificial Intelligence if it is capable of acquiring knowledge in a similar way to a human being, and of establishing new connections between the learned knowledge and existing information.
RankBrain is part of Google Search, which in turn is based on an algorithm called "Hummingbird". It considers up to 200 different factors, called signals, to decide what information is likely to satisfy the search intent of the user and to be displayed in the search results. This information can take any form, be it websites, apps, videos, or documents.
Prior to implementing RankBrain, search results were calculated using statistical, human-specified rating scales.
In the case of unsatisfactory search results, these values were checked and adjusted, sometimes even deleted altogether. These benchmarks can be thought of as filters related to selected aspects of the Google algorithm.
Some of these filters are well known and got names borrowed from the wildlife. For example "Panda", which considers content-related aspects and relevance of websites, "Penguin", which analyzes the quality of backlinks of websites or "Pidgeon" with effects on the local search. (Read more about that in the article " Google Updates") Because site owners try to manipulate these filters with specially targeted content or backlinks, they often need to be updated.
However, only the filter and not the entire Google algorithm will be updated. In this way, high-quality search results can be achieved without affecting the central Google algorithm.
RankBrain, on the other hand, is not based purely on statistical values, but rather abstracts them on the basis of machine learning. It uses machine learning to transform information into mathematical entities (vectors). This makes it possible to compare information, recognize patterns and create new connections. Thus, known and unknown searches can be interpreted and the best matching search result delivered.
The RankBrain system was rolled out for the first time in 2015. Google admits, However, that the company cannott explain exactly how the RankBrain technology ultimately affects. "We have no idea," said Gary Illyes, webmaster trends analyst at Google SMX Advanced conference in 2016 (Source: Search Engine Land )
The use of RankBrain and the transition from purely statistical to semantic-related ranking factors in the Google algorithm leads to the following conclusions:
These can be images, videos, graphics, etc. that deserve separate display in Google's search results (Knowledge Graph, Universal Search, Featured Snippets).
RankBrain is still a long way from becoming an independent AI system, as it does not yet learn by itself , but is "trained" by Google engineers. Nonetheless, the implementation of machine learning in complex search queries should make it easier to answer previously unknown searches without human help.
The behavior patterns of online users are constantly evolving. More ore and more precise information will be found faster and faster with artificial intelligence is expected. As information technology becomes lighter and more mobile friendly.
Even today, the search queries are not only manually, but by voice control - and this with increasing tendency. There are wearables that understand not only language and writing, but also gestures, such as, 'the Dash' of the Munich startup BRAGI .
An artificial intelligence search system could proactively satisfy users' intentions before asking a question, and become a wizard from a search engine.
The question remains, are we always trying to make the changes, instead of writing the story, by trying to make the unknowable writable?
Also Read: Top Five Programming Languages Must Learn In 2018
The fact is artificial intelligence (AI) technology is expanding into ever more increasingly. A driving force behind this development is the Google parent company Alphabet, which inscribes artificial intelligence ever deeper into the DNA of its all areas of life covering subsidiary companies. The company's product portfolio includes medical technology (contact lenses for diabetics), mobility (drone delivery, self-driving cars), renewable energies (wind turbines, so-called "energy kites") and smart home technology (NEST) (Source: All of Google's , Alphabet's-companies and products from A to Z ).
In the everyday life of most people, however, the Google search engine is still the most important point of contact with the company from Mountain View. Whether it's on the desktop, mobile phone or smart TV, Artificial Intelligence and Machine Learning have now made a significant contribution to choosing the right results on billions of searches every day.
This is particularly clear on a part of the Google algorithm, which officially carries the somewhat strange name "RankBrain".
The first official announcements for the use of artificial intelligence under the name RankBrain, there was Google in early 2015.
The system, it was said, help to interpret the search results on unknown, new searches and more complex inputs from the natural voice search correctly and appropriate To deliver search results.
At that time, according to Google, about 15% of all Google searches were affected. Meanwhile, it is active in the background with all around three trillion searches per year.
Machine Learning and Artificial Intelligence
Machine learning, in its purest form also called "Deep Learning", is an automated learning process in which a system can generalize the learning data by pattern recognition, comparisons and examples and apply it to new, unknown situations.
hind RankBrain hides an artificial neural network, which learns without human instruction, from existing data and information assignments and relationships to recognize and use.
But such a system can only be called Artificial Intelligence if it is capable of acquiring knowledge in a similar way to a human being, and of establishing new connections between the learned knowledge and existing information.
RankBrain and Google Search
RankBrain is part of Google Search, which in turn is based on an algorithm called "Hummingbird". It considers up to 200 different factors, called signals, to decide what information is likely to satisfy the search intent of the user and to be displayed in the search results. This information can take any form, be it websites, apps, videos, or documents.
Prior to implementing RankBrain, search results were calculated using statistical, human-specified rating scales.
In the case of unsatisfactory search results, these values were checked and adjusted, sometimes even deleted altogether. These benchmarks can be thought of as filters related to selected aspects of the Google algorithm.
Some of these filters are well known and got names borrowed from the wildlife. For example "Panda", which considers content-related aspects and relevance of websites, "Penguin", which analyzes the quality of backlinks of websites or "Pidgeon" with effects on the local search. (Read more about that in the article " Google Updates") Because site owners try to manipulate these filters with specially targeted content or backlinks, they often need to be updated.
However, only the filter and not the entire Google algorithm will be updated. In this way, high-quality search results can be achieved without affecting the central Google algorithm.
How RankBrain works
RankBrain, on the other hand, is not based purely on statistical values, but rather abstracts them on the basis of machine learning. It uses machine learning to transform information into mathematical entities (vectors). This makes it possible to compare information, recognize patterns and create new connections. Thus, known and unknown searches can be interpreted and the best matching search result delivered.
- In addition, RankBrain can enrich the process of vectoring with contextual data. With its Android operating system or Chrome browser, Google now has access to a lot of contextual information such as a user's location, device, the language they speak, etc.
- The bulk of this information also referred to as big data, is collected by Google in a constant process and updated.
- Thanks to the increasing availability of data, the learning process of RankBrain can be continuously improved. The system is becoming smarter about registering user behavior, recognizing and interpreting relationships.
Facts and application
The RankBrain system was rolled out for the first time in 2015. Google admits, However, that the company cannott explain exactly how the RankBrain technology ultimately affects. "We have no idea," said Gary Illyes, webmaster trends analyst at Google SMX Advanced conference in 2016 (Source: Search Engine Land )
However, the following information is known about RankBrain:
- It is part of the Google algorithm
- RankBrain is developing dynamically and is in a constant learning process
- it is already applied to all searches
- RankBrain does not necessarily affect the search results
- His learning process is based, among other things, on user behavior patterns combined with contextual information
- artificial intelligence does not try to understand individual words, but the phrases and the user intention RankBrain is relevant for the following types of searches:
- voice search
- Unknown searches
- Long-tail searches
- Search queries with ambiguous words
Which websites will benefit from RankBrain?
The use of RankBrain and the transition from purely statistical to semantic-related ranking factors in the Google algorithm leads to the following conclusions:
- Sites with unique content that can satisfy user intent (even at item level)
- Websites with a clear and clear information architecture
- Websites with the ability to satisfy conversational queries
- Entities (search queries around personalities, places and facts) and their semantic contexts
- Structured data (Structured Data Markup)
- User-oriented content
- Diversity of content related to all aspects of an entity
These can be images, videos, graphics, etc. that deserve separate display in Google's search results (Knowledge Graph, Universal Search, Featured Snippets).
RankBrain is still a long way from becoming an independent AI system, as it does not yet learn by itself , but is "trained" by Google engineers. Nonetheless, the implementation of machine learning in complex search queries should make it easier to answer previously unknown searches without human help.
Today and tomorrow
The behavior patterns of online users are constantly evolving. More ore and more precise information will be found faster and faster with artificial intelligence is expected. As information technology becomes lighter and more mobile friendly.
Even today, the search queries are not only manually, but by voice control - and this with increasing tendency. There are wearables that understand not only language and writing, but also gestures, such as, 'the Dash' of the Munich startup BRAGI .
An artificial intelligence search system could proactively satisfy users' intentions before asking a question, and become a wizard from a search engine.
The question remains, are we always trying to make the changes, instead of writing the story, by trying to make the unknowable writable?
Also Read: Top Five Programming Languages Must Learn In 2018
Other sources:
- Bloomberg: Google Turning Its Lucrative Web Search Over to AI Machines. October 2015 http://www.bloomberg.com/news/articles/2015-10-26/google-turning-its-lucrative-web-search-over-to-ai-machines
- Search Engine Land: FAQ: All about the Google RankBrain algorithm. 23.06.2016 http://searchengineland.com/faq-all-about-the-new-google-rankbrain-algorithm-234440
- Search Engine Land: Google uses RankBrain for every search, impacts rankings of "lots" of them. 06/23/2016 http://searchengineland.com/google-loves-rankbrain-uses-for-every-search-252526
- The SEM Post: Google's RankBrain: 9 Industry Experts Weigh In . 27.10.2015 http://www.thesempost.com/googles-rankbrain-9-industry-experts-weigh-in/
- Code Google: Word2vec. 30.07.2013 https://code.google.com/archive/p/word2vec/
Related Posts
Best Microphone for YoutubersHello my dear friends are you all searching on the internet the best microphone for recording the videos and lectures o- JIO PRIME Membership After 31st March 2018 ? New Updates And PlansHello, Jio users are you afraid of losing JIO PRIME Membership After 31st March 2018?Now, most of the people are search
- Top 5 Best Freelancing Websites Must JoinBest Freelancing Websites: Online freelancer sites and marketplaces are altering the way we operate. We can operate fro
- Top 5 Upcoming Technologies in 2018Hello, my dear friends today we are going to share something interesting and Mind-blowing fact about upcoming science t
- How To complete 4k Hours Watchtime And 1k Subscribers EasilyHello my dear friends and dear YouTube creators, as you all know that as per the latest updates of YouTube rules and re
- How to Get Jio Prime Membership For Free April 2019Hey Guys! Are you Jio prime member ? and wants to continue the benefits of Jio prime membership for free? Then you are
Artificiaal intelligence in action or: What is RankBrain?
4/
5
Oleh
NIYAMATULLA
