Matt Cutts on the future search technology
Read / WriteWeb recently published an interview with Matt Cutts , the head of Google’s anti-spam department. Mat shared his thoughts on the technology of finding the future.
- Tell us about the work of Google on the next generation search engine.
- The main innovation of the search engine of the future is the possibility of personalized search when the user receives only relevant and necessary information. That is, the person enters a simple request, and Google takes care of the details. For example, a New Zealand resident enters the word "Bank" and receives a list of banks only from his country. This is the most primitive selection, by location, the idea is easy to develop, it has great potential.
')
- Recently, the personalization service is available through Google Accounts. Is this the first step to personalized search?
- Yes, the user does not need to adjust the search parameters every time. While he is in the system, she adapts the search results for him. If this function turns out to be superfluous, the user is always entitled to click on the link in the upper right corner and abandon it, having logged out.
- What about the semantics of the language? How important is it to understand human natural language for search technology, and how far has Google come in this direction?
- We use different technologies and semantics of the language as well. Semantics is attempted to be implemented in one form or another by all existing search engines, but one should not bet on any one technology. In this case, Google’s approach is very practical, and can be traced in an existing search engine.
- Now there are many alternative technologies for searching information in the network. Which of them do you consider the most promising?
- The already mentioned personalized search technology and the idea of artificial intelligence seem to be the most natural. Also worth mentioning is the visualization, restructuring of the entered query. Sometimes at the bottom of the page, along with the search results, you are prompted to search by automatically restructured or related query, which, according to the search engine, will allow you to get better search results. This mechanism Google has been testing for a long time, but we don’t want to run it completely until we debug it to a sufficiently high quality level.
- At the end of last year, an experimental search site SearchMash was launched, which has many innovative features and is built using Ajax technology. Is this a prototype of the new version of google.com?
- There is a possibility, of course, but it’s far from the fact that all the features of SearchMash will be implemented in the Google search engine. You always have to compromise because of the compatibility between different browsers, user needs, the size of the widget, and the time it takes to implement the function. For example, in SearchMash, you can start typing anywhere on the page, but the text will appear in the query window - a useful feature that will not work in all browsers. SearchMash is a kind of field for experimenting with the user interface. We can try out different approaches and see how consumers respond to them. All technologies undergo long-term testing before being implemented in google.com.
- Do you have plans to introduce the function of specialized search for blogs, technical resources, etc.?
- I propose to recall the already implemented search functions of patents, source codes of program code, books and other resources of a fairly narrow specialization. Google Calendar searches in calendar entries, Gmail searches emails. In addition, there is Google Custom Search Engine (CSE), a technology that allows you to actually build your own search engine that searches for specific resources. The user can easily add 5000 addresses and then filter between them. Thus, if a person wants to find, for example, a number of podcasts on technical topics, you can use search engines already configured using CSE technology or create your own. I think this is a very correct approach, and it is very popular.
- You participated in the creation of spam filtering technology. Tell us how Google manages to highlight web spam in the background of search results?
- Over the past year we have seriously improved our anti-spam filters and even created a special measurement system that shows how far we have advanced. Last year we paid a lot of attention to foreign language spam, so that the search in any language would be relevant.
- Last year, Google acquired a YouTube resource and launched Google Video, as well as a video clip search service. What directions does the company follow in the field of video search now?
- The video itself is more interesting and informative than text, pictures or podcasts, because it also contains a video sequence. At the same time, it is much more difficult to index and add to the search base. We use a rating system, pay attention to the number of sites that link to a video and to the site containing it. Thus, search results are usually satisfactory.
Source: Read / WriteWeb
- Tell us about the work of Google on the next generation search engine.
- The main innovation of the search engine of the future is the possibility of personalized search when the user receives only relevant and necessary information. That is, the person enters a simple request, and Google takes care of the details. For example, a New Zealand resident enters the word "Bank" and receives a list of banks only from his country. This is the most primitive selection, by location, the idea is easy to develop, it has great potential.
')
- Recently, the personalization service is available through Google Accounts. Is this the first step to personalized search?
- Yes, the user does not need to adjust the search parameters every time. While he is in the system, she adapts the search results for him. If this function turns out to be superfluous, the user is always entitled to click on the link in the upper right corner and abandon it, having logged out.
- What about the semantics of the language? How important is it to understand human natural language for search technology, and how far has Google come in this direction?
- We use different technologies and semantics of the language as well. Semantics is attempted to be implemented in one form or another by all existing search engines, but one should not bet on any one technology. In this case, Google’s approach is very practical, and can be traced in an existing search engine.
- Now there are many alternative technologies for searching information in the network. Which of them do you consider the most promising?
- The already mentioned personalized search technology and the idea of artificial intelligence seem to be the most natural. Also worth mentioning is the visualization, restructuring of the entered query. Sometimes at the bottom of the page, along with the search results, you are prompted to search by automatically restructured or related query, which, according to the search engine, will allow you to get better search results. This mechanism Google has been testing for a long time, but we don’t want to run it completely until we debug it to a sufficiently high quality level.
- At the end of last year, an experimental search site SearchMash was launched, which has many innovative features and is built using Ajax technology. Is this a prototype of the new version of google.com?
- There is a possibility, of course, but it’s far from the fact that all the features of SearchMash will be implemented in the Google search engine. You always have to compromise because of the compatibility between different browsers, user needs, the size of the widget, and the time it takes to implement the function. For example, in SearchMash, you can start typing anywhere on the page, but the text will appear in the query window - a useful feature that will not work in all browsers. SearchMash is a kind of field for experimenting with the user interface. We can try out different approaches and see how consumers respond to them. All technologies undergo long-term testing before being implemented in google.com.
- Do you have plans to introduce the function of specialized search for blogs, technical resources, etc.?
- I propose to recall the already implemented search functions of patents, source codes of program code, books and other resources of a fairly narrow specialization. Google Calendar searches in calendar entries, Gmail searches emails. In addition, there is Google Custom Search Engine (CSE), a technology that allows you to actually build your own search engine that searches for specific resources. The user can easily add 5000 addresses and then filter between them. Thus, if a person wants to find, for example, a number of podcasts on technical topics, you can use search engines already configured using CSE technology or create your own. I think this is a very correct approach, and it is very popular.
- You participated in the creation of spam filtering technology. Tell us how Google manages to highlight web spam in the background of search results?
- Over the past year we have seriously improved our anti-spam filters and even created a special measurement system that shows how far we have advanced. Last year we paid a lot of attention to foreign language spam, so that the search in any language would be relevant.
- Last year, Google acquired a YouTube resource and launched Google Video, as well as a video clip search service. What directions does the company follow in the field of video search now?
- The video itself is more interesting and informative than text, pictures or podcasts, because it also contains a video sequence. At the same time, it is much more difficult to index and add to the search base. We use a rating system, pay attention to the number of sites that link to a video and to the site containing it. Thus, search results are usually satisfactory.
Source: Read / WriteWeb
Source: https://habr.com/ru/post/5417/
All Articles