"Yandex" has launched a new "smart" search. How to use it? Yandex launched a new version of the search engine What's new in the Yandex search engine

Yandex again pleases users with innovations to improve the quality of information search. Today, the changes have affected snippets - a "Read more" button and a chat with the company have appeared. Let's take a closer look at these updates.

A new “Read more” button has appeared in the desktop version of Yandex search, with its help the user can see an extended snippet of the site. When you click on the button, the explanatory text associated with the request appears.

Previously, this feature was available only for mobile devices, but now it has been introduced into the functionality of the desktop version of search results on an ongoing basis.

In the recent past, a Yandex snippet displayed a limited amount of information - 240 characters with spaces, about 3 lines. Often the user did not have enough of this information, and there was a need to go to the site in search of information of interest.

Now, with the addition of this button, the process of searching for information for the user can speed up several times.

View of the new Yandex snippet

What the Yandex snippet looked like before:

Now the Yandex snippet in a collapsed state looks like this

The snippet looks like this in the mobile version

Appropriate view in mobile issuance.

Pros and cons of the Yandex rich snippet

Based on this information, the user can make a more informed choice about the usefulness of the resource. And for the site owner, a featured snippet is an opportunity to get more targeted traffic, improve behavioral factors - for example, reduce bounce rate and increase conversion traffic.

However, the introduction of this function may also have a negative impact on the search traffic of the site, since this description will be quite enough for the user to get the information they are looking for without going to the site.

As for the length of the featured snippet, Yandex experts say this:

“The principle is this: the featured snippet can be no more than 3 times longer than the current one”. That is, the maximum length of the snippet at the moment can be up to 9 lines or 650-700 characters.

“Regarding where extended descriptions come from, the same rules apply here as for regular snippets,” Yandex experts explain. The search engine generates a snippet in accordance with the user's request, and if the information from the description meta tag is not enough, Yandex will take the information from the site page.

You can read more about the rules for Yandex snippets in this article.

You should be careful, there is a chance that service information will get into the extended snippet: text from pagination buttons, filters, sorting, and more, so you should close it in the code with tags.

Jivosite chat in Yandex search results

Also, a chat appeared in the Yandex snippet. Yandex together with Jivosite added a chat to the search results page. Starting April 26, 2018, customers can contact support for a site that has Jivosite Chat installed without going to the site.

After connecting the chat, you will be able to:

• enable or disable chat in the search;
• add an automatic reply;
• add hints;
• add icon;
• specify the organization's working hours and limit the number of pages on which the chat will be shown.

Instructions for adding a chat to your site are presented on the official website of the Jivosite service.

Chat view in search results

What the “Chat with Company” snippet looks like in the desktop version:

Please note that there are different icons for starting a chat for the desktop and mobile versions.

After clicking on the "Chat with the company" button, a form for communicating with the site's support appears directly on the search results page.

In the mobile version of the search, the chat window opens on top of the current window, but not in a new tab.

An interesting feature of the chat is the preservation of the history of correspondence of an authorized user with the support of the site on the other end of Jivosite.

After sending the letter, the user has the opportunity to minimize the chat and continue searching for information of interest:

After the site support responds to the user, a new message notification will appear on the right:

Pros and cons of chat

This snippet allows you to make the user's communication with the organization more accessible. The user will be able to make a purchase or ask a question of interest without going to the site.

The downside is that if the operator is not online and the customer has sent a message, the operator will only receive the message when the customer is online next time.

Also, the downside is that the chat on the search results page is not synchronized with the chat on the site itself. Imagine a situation: a user a couple of days ago wrote in support of the site directly from the search results, and a week later decided to go to the site and clarify information about his issue on the site itself. As a result, the operator may not understand what is happening and what the issue is at all. An awkward situation, isn't it? The client will have to spend time re-explaining the issue, or he, not wanting to duplicate the appeal, will completely end the communication. At the moment, there is no solution to this situation, let's hope that in the future, Yandex specialists, together with Jivosite, will find a way out.

Conclusion

Yandex does not stand still and constantly comes up with something interesting. This time useful functions have been added, snippets in Yandex have acquired a new look and new features. This will help improve the quality of information in search results.

If you do not have time, you can always contact us for help in installing this tool.

Yandex launched a new version of the search. It is based on the search algorithm "Korolev". The algorithm uses a neural network to match the meaning of requests and web pages - this allows Yandex to more accurately respond to complex queries. Search statistics and ratings from millions of people are used to train the new version of the search. Thus, not only developers, but also all Yandex users contribute to the development of search.

Words and meanings

Before talking about the present and future of search, let's remember its past. The first search engines appeared in the mid-1990s, when the Internet was very small - the number of sites went into the thousands. To help a person find the right one, it was enough to make a list of web pages where there are words from the search query. There was no talk of complex ranking - that is, ordering pages according to the degree of compliance with the request. It was believed that the more often the words from the query occur in the document, the better it fits.

The Internet grew rapidly and additional selection criteria were needed. Search engines began to take into account links to documents, learned to determine the region where the request came from, and began to pay attention to user behavior.

At some point, there were so many ranking factors - signs by which you can determine how well a page responds to a request - that it became clear that it was impossible to write them all in the form of instructions. It is better to teach the machine to make decisions on its own: what features to use and how to combine them. Yandex came up with Matrixnet for these purposes. This is the machine learning method that builds our ranking formula.

Search, however, still relies on words. Before launching a complex ranking formula, search engines compile a list of "pre-qualified" web pages—those that contain the words in the query. We humans understand that the same meaning can be expressed in different words. A web page may not contain all of the words in a query, but still respond very well to it. However, it is quite difficult to explain this to a machine.

Yandex took the first step towards search by meaning last year when the company introduced the Palekh search algorithm. It is based on a neural network. Neural networks excel at tasks that humans have traditionally done better than machines, such as recognizing speech or objects in images.

By launching Palekh, the company taught the neural network to convert search queries and web page titles into groups of numbers - semantic vectors. An important property of such vectors is that they can be compared with each other: the stronger the similarity, the closer the request and the header are in meaning to each other.

How the Korolev algorithm works

The Korolev search algorithm compares the semantic vectors of search queries and entire web pages - not just their titles. This allows you to reach a new level of understanding of meaning. Imagine that you first heard about Leo Tolstoy's novel War and Peace. Of course, you can make sense of the title - for example, suggest that there are many battle scenes in the book. But in order to learn all the intricacies of the plot and give comprehensive answers to questions about the novel, you will need to read it in its entirety.

As in the case of Palekh, the texts of web pages are converted into semantic vectors by a neural network. This operation requires a lot of computing resources. Compare: it will take you seconds to read the title of a book, but it will take you hours, days or even weeks to read it all from cover to cover. Therefore, Korolev does not calculate page vectors in real time, but in advance, at the indexing stage. When a person makes a request, the algorithm compares the request vector with the page vectors it already knows.

Such a scheme allows you to start the selection of web pages that match the query in meaning at the early stages of ranking. In Palekh, semantic analysis is one of the final stages: only 150 documents go through it. In Korolev, it is produced for 200 thousand documents - that is, more than a thousand times more. In addition, the new algorithm not only compares the text of a web page with a search query, but also pays attention to other queries that people come to this page for. In this way, additional semantic connections can be established.

People teach machines

Yandex believes that the use of machine learning, and especially neural networks, will sooner or later teach search to operate with meanings at the human level. But without the help of people, this is not possible. In order for a machine to understand how to solve a particular problem, it is necessary to show it a huge number of examples: positive and negative. Such examples are given by Yandex users.

The neural network used by the Korolev algorithm is trained on impersonal search statistics. Statistics collection systems take into account which pages users go to for certain queries and how much time they spend there. If a person opened a web page and "hung" there for a long time, he probably found what he was looking for - that is, the page responds well to his request. This is a positive example. It is much easier to pick up negative examples: just take a request and any random web page.

Matrixnet, which builds a ranking formula, also needs people's help. For search to evolve, people must constantly evaluate its work. Once upon a time, only Yandex employees, the so-called assessors, were engaged in grading. But the more ratings, the better - so we decided to involve everyone in this and launched the Yandex.Toloka service. Now more than a million users are registered there: they analyze the quality of search and participate in improving other Yandex services. Tasks on Toloka are paid - the amount that can be earned is indicated next to the task. For more than two years of the existence of the service, tolokers have given about two billion ratings.

Modern search is based on complex algorithms. Algorithms are invented by developers, and taught by millions of Yandex users. Any request is an anonymous signal that helps the machine understand people better. Therefore, Yandex will not be mistaken if it says: a new search is a search that we did together.

This week, August 22, Yandex launched a new version of the search with the "Korolev" algorithm. It is based on a neural network that allows it to match the meaning of a request and a web page and respond much more accurately to complex and ambiguous requests. To train a new version of the search, search statistics and estimates of millions of people are used: it turns out that not only developers, but all users in general, contribute to the development of the system.
The presentation of the "Korolyov" took place, which is symbolic, in the Moscow Planetarium. Andrey Styskin, Head of Yandex.Search, Alexander Safronov, Head of Yandex.Search Relevance Service and Olga Megorskaya, Head of Yandex.Search Data Processing Department performed on stage.

From Matrixnet to neural networks

Yandex came up with Matrixnet to process requests - a machine learning method that was used to build the author's ranking formula. However, the search continued to rely on words. But what about queries that users formulate allegorically or associatively? Then the web page you are looking for does not have to contain strictly all the words from the query. But how do you explain this to a machine? If only she understood us as a person...

In the end, scientists came up with something at the intersection of technology and biology - an artificial neural network (ANN). According to the wording of Wikipedia, this is "a mathematical model, as well as its software or hardware implementation, built on the principle of organization and functioning of biological neural networks - networks of nerve cells of a living organism." Neural networks are able to process information like we do and, most importantly, learn and hone skills like living beings. Actually, they are the basis of a full-fledged artificial intelligence, the appearance of which is a matter of time.

Last year, Yandex introduced the Palekh search algorithm based on a neural network. He showed excellent results in solving problems that were usually only possible for people: he did an excellent job of recognizing speech and objects in images. "Palekh" has learned to convert search queries and web page titles into groups of numbers - semantic vectors. Their important property is that vectors can be compared with each other: the stronger the similarity, the closer the query and the title are in meaning.

"Kings". who understands

The neural network learns like a child. To master this, she needed a huge number of examples. Actually, all users of the service were engaged in spontaneous training of Korolev in one way or another: search statistics and estimates of millions of people were used. Yandex is gradually learning to more and more accurately recognize semantic connections, like: [a picture where the sky is twisting] is about a Van Gogh painting, [a lazy cat
from Mongolia] - manul.

Search is a very complex system. Thousands of engineers are working to ensure that she understands a person and helps to solve his problems. In Korolyov, we have combined machine intelligence and the efforts of millions of people. Our users improve search with us by asking questions and helping to train our algorithms.
Andrey Styskin, Head of Yandex Search.

“Modern search is based on complex algorithms. Algorithms are invented by developers, and taught by millions of Yandex users. Any request is an anonymous signal that helps the machine understand people better. Therefore, we will not be mistaken if we say: the new search is the search that we did together,” the Yandex blog post reads.

In the more than two-year history of Yandex.Toloka, the most productive and diligent participant has been identified. They became Ilya Mikhalenko from Chelyabinsk. The guy came to the presentation of "Korolev" in Moscow to receive a well-deserved award from the hands of the search engine team.

New search in action

Image search has been greatly improved. With them, as a rule, there is always some kind of "hell": the search engine either thoughtlessly gives out all the images in the name of which the words from the query are used, or takes into account the text of the article that the picture illustrates. If you are looking for something that would meet the vague needs of the soul, then get ready to be disappointed. "Korolev" analyzes exactly what is shown in the picture, therefore it is able to please with a non-trivial approach.

As an example, tests were given not the most obvious request - [a cat in space]. Dogs were in orbit quite often, but the mustachioed-striped disciplined conquerors of space did not work out. Only one attempt is known for certain: in 1963, the French launched the cat Felicette into suborbital flight. Romantic, but short-sighted, - as soon as the scientists opened the hatch of the landing capsule, the murk was like that. The photo session did not take place.

Upon request, the search engine gives out not only little animals in spacesuits and surreal photo-toads, but a photo of a cat in a washing machine, which is quite similar to the hatch of a spaceship. But this is not stated in the description.

For the solemn launch of the new search engine, the entire Yandex.Search team took the stage. A little countdown and... Let's go! Now everyone can experience the capabilities of the insightful "Queen". The main thing is that its current capabilities are not static, but are in constant development.

To end the evening, the organizers have prepared something completely unexpected - a communication session with real astronauts from orbit. They personally responded to some of the popular search engine user queries about space and answered questions from those present.

The neural network analyzes not only the title, but the entire page, while the search engine determines its essence even at the indexing stage

MOSCOW, 22 August. /TASS/. "Yandex" has launched a new version of the search, which is based on a comparison of the meaning of the request and the web page, the company said. The new version works on the "Korolev" algorithm, which uses a neural network to determine what exactly the user needs. The neural network analyzes not only the title, but the entire page, while Yandex determines the essence of the page in advance, at the indexing stage.

Another feature of the "Queen" is that it also takes into account the meaning of other requests by which people switch to it. “In order for a neural network to evaluate the semantic similarity of a query and a document, it needs a huge number of examples. Such examples are given by impersonal search statistics: what sites people go to for queries and how much time they spend there. So, if a person went to the page and looked through it for a while, most likely it is close in meaning to the request. Using the search statistics of millions of people, Yandex is learning to understand semantic connections. For example, he will understand that in the query “a picture where the sky swirls”, it is about a Van Gogh painting, and in the query “a lazy cat from Mongolia” it is about a manul,” the company said in a press release.

Last year, Yandex already launched a system based on neural networks - Palekh. The previous system indexed 150 pages, in the "Queen" the number of pages that the search compares in meaning with the query has grown to 200 thousand.

The new algorithm was named after the founder of Russian cosmonautics Sergei Korolev.

“And today we are launching a new ranking algorithm for the Queens. Why did we choose this name? Sergei Pavlovich Korolev fulfilled the dream of mankind to fly into space. For us at Yandex, today's launch is just as important a technological breakthrough towards the dream of a search that understands users,” Alexander Safronov, head of the Yandex Linguistics Relevance Service, said at the presentation of the new version of the search.

Hall help

To train a search engine, you need to evaluate the quality of responses. Previously, "Yandex" evaluated the quality of the search with the help of its evaluators. The new search will take into account the ratings that users of the Yandex.Toloka service, a distributed network of appraisers, will put up. The service allows anyone to complete tasks and receive rewards for them, at the moment it has more than a million registered users. Anyone can register on the platform.

Yandex is the largest search engine in Russia. The company's share in the Russian search market (including search on mobile devices) in the second quarter of 2017 averaged 54.3%, in the first quarter of this year - 54.7% (according to the Yandex.Radar analytical service). According to Liveinternet.ru, in June of this year, the search share of Yandex was 51.3%.