The Atom platform from Yandex - the Internet adapted for everyone
Recently, at YaC 2013, Yandex announced a new Atom platform. It uses a variety of technologies that were created for different tasks, to solve one big thing - to change the Internet so that each of us ceased to be an abstraction for him, and became a person with his own character and interests.
Gradually, everyone came to understand that the Internet is determined not by documents, but by people. It is connected with reality and consists of the needs, preferences and tasks of people - like a world of atoms. The Atom program is about a person, she puts forward and tests hypotheses about what he wants, what he wonders, what he needs on a particular site.
')
Each step in the development of our technologies was necessary in order to cope with the growth of content on the Internet and be able to show people the answers that are suitable for him. In 2009, we launched the search platform Arzamas . Yandex began to organize search results using the ranking formula for a particular region. At the same time, we have implemented our own machine learning method, Matrixnet , which is now used not only in search technologies in Yandex. For example, it is used by Seismotech for processing seismic data and CERN for analyzing the decay of mesons . In 2011, Crypt appeared, which was also created on the basis of Matrixnet. It is used in advertising technologies, automatically calculating the sex, age and other characteristics of a person.
But not all the interests of people can be described in the questionnaire format, so we began to engage in the personalization of the issue, adapting to the short-term and long-term interests of the user. To this end, the search platforms Kaliningrad and Dublin were launched. These are actually technologies for adapting a resource for a user, which we apply primarily in the search, changing the output for a specific person who made the request.
Adapting the Internet to the user's interests is a natural process. And sites began to face problems similar to those that arose with us. What to show the user when a lot of content? What recommendation should be given when the content on each site is about the same? People still choose services that solve their problem faster, and remember those that gave a good recommendation.
It is good if the site has at least geotargeting and is adapted to the region of the user. Some are beginning to calculate the interests of the user, clearly asking about them. For example, sports.ru asks each new user a question about their favorite football team.
The most difficult step that most still have to do is work with a large number of factors and fuzzy signals about interests. The market is already approaching the need for this step. The Atom program, which uses big data processing and machine learning technologies, can help make it.
We can realize such things, thanks to a combination of two rather unique factors. On the one hand, we have a huge amount of knowledge about users and their behavior on the Internet. On the other - the ability to store and process large data in real time . We have created these technologies for ourselves, because without them the search company cannot make a quality product. And their combination is probably five companies in the world. We are the only ones who are ready to declare the desire to make such technologies an accessible product.
Why do we announce our program without telling how we applied it on our services? If we did it for ourselves, it is obvious that we would announce it at another stage of readiness. But we are talking about the thing, the rules of existence of which need to be determined by all together.
The discovery of the Atom is somewhat similar to the discovery of the Islands . While working on integrating the responses of services into Search, at some point we realized that solving the problems of users together with the site owners is more promising. This is how the Islands appeared. When we worked on personalization, it came to an understanding that it was impossible to make a tool for solving problems personalized, and leave the solutions and services on the Internet not personalized. That is how Atom appeared. Today we tell you about Atom in order to start doing it together.
Both the Islands and Atom are based on an understanding of changes on the Internet: it is increasingly connected with reality, with the tasks of people. The transition from documents to services has already happened. Islands come to this on the one hand - they accompany the person on the task. Atom - on the other: it will allow the Internet to work with a specific person, and not with abstraction. Obviously, this helps to solve his tasks, and in a broad sense: both transactional and those in which a person is primarily interested in content.
In both products, we fundamentally change the approach of creation. As a metaphor, we use crowdsourcing all the time: the scale of change is too large and in order for everything to work well, organically, the ecosystem needs the contribution of all its actors, all participants in the process.
The use of our technologies by other sites will allow us to make the resources more appropriate for an individual - so that they become more convenient and more interesting for him. Talking about Atom at YaC, we proposed several basic scenarios for which such adaptation could be made: changing the start page to a user category, ranking objects within a category, smart recommendations based on a combination of user-based ranking and knowledge of the subject area to which the site is dedicated. This is something like anchor points - the way they develop will depend a lot on what webmasters working with sites will see in Atom.
The result that we see - moreover, it seems inevitable - is the ability to adapt to the specific person who has come to the resource that has appeared in the entire Internet. It will bring more convenience to the user and will give interesting content to him. The site, respectively, will be able to increase the conversion and depth of viewing. If you are talking about transactions, the person will spend less clicks getting to the goal. If you have a content site - the person does not see boring materials on it, is more likely to stick, with much more - it will stick to it for a long time. If you are a smart store, then thanks to Atom you will get more and give more to the user: he will quickly find the phone that suits him, and in the recommended products he will see an unobvious, but cute case.
Today it is strange to see a large site, incorrectly displayed in different browsers. And ten years ago it was common. In the end, it is strange to see a service that does not support login through social networks - this was not even five years ago. Three or four years, and the person simply becomes confused, not having met his usual level of convenience on the next site.
Thanks to Atom, everyone will be able to see content that, other things being equal, would pass by it. The recommendatory component will “pull out” for a person more than what he himself has already searched for and found on the site: this does not narrow, but will expand his interests. The user will see only those products or the information that they need, are interesting and useful, and for which he is willing to spend money or time.
But it is hardly reasonable to assume that the person will be interested in all the new information. We will show him the one that could potentially be of interest to him. This is especially evident in examples with objects from the real world: a person was not going to find some kind of content, and we realized that it would be right for him to show it. An atom does not close a person in what he already knows, but opens something from the unknown that he will like.
It is impossible to show the person all the content of the world, and it is useless: the housewife does not want to read about Warcraft, and the sysadmin - about rhinestones. But within the framework of what a person does not see, there is something that he would be more or less interested in. We help lift up what has a chance to catch him - and get out of the current bubble.
If you want to expand the sphere of interests where it will be understood and accepted by man, Atom will help you. If you think you can convince a person of the need for what he really doesn’t need, or lead him to where he really doesn’t need, you can hope for zombies, a whip, or other strange tools.
All the technologies we share will be based on Yandex’s key expertise - working with data: machine learning , data extraction, factor creation , evaluation of the statistical reliability of the results, an experiment pipeline.
Currently, technology is being tested on several Yandex services and with a number of external partners. The number of external testing will be increased, and when we get stable results and understand that the technology is ready for scaling, we will return to you with well-developed rules of participation. If you are ready to share your vision of using Atom or to connect to our experiments, write to us at atom-experiments@yandex-team.ru.
Gradually, everyone came to understand that the Internet is determined not by documents, but by people. It is connected with reality and consists of the needs, preferences and tasks of people - like a world of atoms. The Atom program is about a person, she puts forward and tests hypotheses about what he wants, what he wonders, what he needs on a particular site.
')
Each step in the development of our technologies was necessary in order to cope with the growth of content on the Internet and be able to show people the answers that are suitable for him. In 2009, we launched the search platform Arzamas . Yandex began to organize search results using the ranking formula for a particular region. At the same time, we have implemented our own machine learning method, Matrixnet , which is now used not only in search technologies in Yandex. For example, it is used by Seismotech for processing seismic data and CERN for analyzing the decay of mesons . In 2011, Crypt appeared, which was also created on the basis of Matrixnet. It is used in advertising technologies, automatically calculating the sex, age and other characteristics of a person.
But not all the interests of people can be described in the questionnaire format, so we began to engage in the personalization of the issue, adapting to the short-term and long-term interests of the user. To this end, the search platforms Kaliningrad and Dublin were launched. These are actually technologies for adapting a resource for a user, which we apply primarily in the search, changing the output for a specific person who made the request.
Adapting the Internet to the user's interests is a natural process. And sites began to face problems similar to those that arose with us. What to show the user when a lot of content? What recommendation should be given when the content on each site is about the same? People still choose services that solve their problem faster, and remember those that gave a good recommendation.
It is good if the site has at least geotargeting and is adapted to the region of the user. Some are beginning to calculate the interests of the user, clearly asking about them. For example, sports.ru asks each new user a question about their favorite football team.
The most difficult step that most still have to do is work with a large number of factors and fuzzy signals about interests. The market is already approaching the need for this step. The Atom program, which uses big data processing and machine learning technologies, can help make it.
We can realize such things, thanks to a combination of two rather unique factors. On the one hand, we have a huge amount of knowledge about users and their behavior on the Internet. On the other - the ability to store and process large data in real time . We have created these technologies for ourselves, because without them the search company cannot make a quality product. And their combination is probably five companies in the world. We are the only ones who are ready to declare the desire to make such technologies an accessible product.
Why do we announce our program without telling how we applied it on our services? If we did it for ourselves, it is obvious that we would announce it at another stage of readiness. But we are talking about the thing, the rules of existence of which need to be determined by all together.
The discovery of the Atom is somewhat similar to the discovery of the Islands . While working on integrating the responses of services into Search, at some point we realized that solving the problems of users together with the site owners is more promising. This is how the Islands appeared. When we worked on personalization, it came to an understanding that it was impossible to make a tool for solving problems personalized, and leave the solutions and services on the Internet not personalized. That is how Atom appeared. Today we tell you about Atom in order to start doing it together.
Both the Islands and Atom are based on an understanding of changes on the Internet: it is increasingly connected with reality, with the tasks of people. The transition from documents to services has already happened. Islands come to this on the one hand - they accompany the person on the task. Atom - on the other: it will allow the Internet to work with a specific person, and not with abstraction. Obviously, this helps to solve his tasks, and in a broad sense: both transactional and those in which a person is primarily interested in content.
In both products, we fundamentally change the approach of creation. As a metaphor, we use crowdsourcing all the time: the scale of change is too large and in order for everything to work well, organically, the ecosystem needs the contribution of all its actors, all participants in the process.
The use of our technologies by other sites will allow us to make the resources more appropriate for an individual - so that they become more convenient and more interesting for him. Talking about Atom at YaC, we proposed several basic scenarios for which such adaptation could be made: changing the start page to a user category, ranking objects within a category, smart recommendations based on a combination of user-based ranking and knowledge of the subject area to which the site is dedicated. This is something like anchor points - the way they develop will depend a lot on what webmasters working with sites will see in Atom.
The result that we see - moreover, it seems inevitable - is the ability to adapt to the specific person who has come to the resource that has appeared in the entire Internet. It will bring more convenience to the user and will give interesting content to him. The site, respectively, will be able to increase the conversion and depth of viewing. If you are talking about transactions, the person will spend less clicks getting to the goal. If you have a content site - the person does not see boring materials on it, is more likely to stick, with much more - it will stick to it for a long time. If you are a smart store, then thanks to Atom you will get more and give more to the user: he will quickly find the phone that suits him, and in the recommended products he will see an unobvious, but cute case.
Today it is strange to see a large site, incorrectly displayed in different browsers. And ten years ago it was common. In the end, it is strange to see a service that does not support login through social networks - this was not even five years ago. Three or four years, and the person simply becomes confused, not having met his usual level of convenience on the next site.
Thanks to Atom, everyone will be able to see content that, other things being equal, would pass by it. The recommendatory component will “pull out” for a person more than what he himself has already searched for and found on the site: this does not narrow, but will expand his interests. The user will see only those products or the information that they need, are interesting and useful, and for which he is willing to spend money or time.
But it is hardly reasonable to assume that the person will be interested in all the new information. We will show him the one that could potentially be of interest to him. This is especially evident in examples with objects from the real world: a person was not going to find some kind of content, and we realized that it would be right for him to show it. An atom does not close a person in what he already knows, but opens something from the unknown that he will like.
It is impossible to show the person all the content of the world, and it is useless: the housewife does not want to read about Warcraft, and the sysadmin - about rhinestones. But within the framework of what a person does not see, there is something that he would be more or less interested in. We help lift up what has a chance to catch him - and get out of the current bubble.
If you want to expand the sphere of interests where it will be understood and accepted by man, Atom will help you. If you think you can convince a person of the need for what he really doesn’t need, or lead him to where he really doesn’t need, you can hope for zombies, a whip, or other strange tools.
All the technologies we share will be based on Yandex’s key expertise - working with data: machine learning , data extraction, factor creation , evaluation of the statistical reliability of the results, an experiment pipeline.
Currently, technology is being tested on several Yandex services and with a number of external partners. The number of external testing will be increased, and when we get stable results and understand that the technology is ready for scaling, we will return to you with well-developed rules of participation. If you are ready to share your vision of using Atom or to connect to our experiments, write to us at atom-experiments@yandex-team.ru.
Source: https://habr.com/ru/post/195982/
All Articles