Audio search, audio SEO and help for podcasts - Google experts are already working on this

According to Podcast Insights, the number of US residents listening to podcasts on a monthly basis is has increased more than twice in the last 5 years. Therefore, Google decided to think about implementing a full-fledged search for such audio recordings. Due to this approach, the IT giant plans to double the number of listeners of podcasts not only in the US, but also all over the world.
 
 
More details about this news and possible plans of the company, we will tell below.
 
 
Audio search, audio SEO and help for podcasts - Google experts are already working on this

 
...
+ 0 -

How to speed up mobile search twice. Lecture of Yandex

How to speed up mobile search twice. Lecture of YandexOn the phone, web pages often load longer than on a desktop. The developer Ivan Khvatov tells about the reasons for the backlog and how to deal with it. The lecture consists of several parts: the first - about the main stages of loading the page on mobile devices, the second - about the techniques that we use to accelerate the load, the third - about our method of adapting the layout for different speeds.
 
 

 
- Hello everyone, my name is Ivan Khvatov, I work in the search infrastructure. Last time I'm working on speeding up the download of search results. I work with layout, backend ...
+ 0 -

How to develop a system that recognizes a person by the keyboard handwriting

We are engaged in consumer lending online. The issue of preventing fraud for us is one of the main. Often, potential scammers think that the security systems of fi nite companies are lower than in traditional financial organizations. But this is a myth. All exactly the opposite.
 
How to develop a system that recognizes a person by the keyboard handwriting
 
In order to reduce the level of potential fraud to zero, we have developed a system of behavioral biometrics that is able to identify a person by the keyboard handwriting. The pilot version was introduced in Russia in August 2017.
 
How it works
 
The system allows you to authenticate the user according to the behavior ...
+ 0 -

Elastic opens the source code for X-Pack

Note. translator. X-Pack is a proprietary extension for ELK products.
 
From open communications to free software, openness has always been in the spirit of Elastic.
 
Therefore, we open the proprietary code X-Pack. X-Pack includes functionality related to security, alerts, monitoring, as well as reports, analytics and machine learning.
 
Why do we do this?
 
Initially, we created X-Pack as a set of proprietary functionality that extends the Elastic stack - Elasticsearch, Kibana, Beats and Logstash. Some functions, for example, monitoring were free. Some, for example, security and machine training ...
+ 0 -

Search the site with Reindexer - it's easy. Or how to make "instant search" all around Habrahabr-y

Hello everyone,
 
In the previous article I wrote that we made a new in-memory DB - fast and with rich functionality - Reindexer .
 
In this article I want to tell you how to use Reindexer to implement full-text search on the site, writing a minimum of application code.
 
Search the site with Reindexer - it's easy. Or how to make "instant search" all around Habrahabr-y
 
In general, full-text search on the site - this is an important feature, in our time, mandatory, for any Internet site. The quality and speed of the search depends on
 
How quickly users will find the information they are interested in or the goods that they plan to purchase.
http://habr-demo.reindexer.org/
 
If w...
+ 0 -

As in hh.ru they test the search for vacancies

I already shared with the story of about our experience of using artificial intelligence in the search for hh.ru, and today I would like to dwell on the measurement of the quality of this search in more detail.
 
As in hh.ru they test the search for vacancies  
For the normal operation of the search system of metrics - local, A /B-tests, queues on the pass, etc., is extremely important, and this system requires separate attention and resources. It's not right to think that it's enough just to cut a steep ML and screw all these metrics with "scotch tape"; It is also not enough to measure the quality of the work of an already running system - it ...
+ 0 -

SSO and Kibana: integration of Kibana with built-in Windows authentication (Single Sign-On)

In this article, I would like to share a way to configure single sign-on (SSO) technology in Elastic Stack, which uses X-Pack to authenticate users and differentiate access to data.
 
SSO and Kibana: integration of Kibana with built-in Windows authentication (Single Sign-On)
Integrating third party Auth with Kibana . As always, a simple and working example was given in such blogs, which is suitable for "home use", but not very applicable in enterprises with complex infrastructure. And only after the release of Elastic Stack v.5.6.? and more specifically - patch in X-Pack 5.6.? which finally made it possible ...
+ 0 -

Two models are better than one. Yandex.Translator's experience

Two models are better than one. Yandex.Translator's experienceWe used to be already told about how machine translation appeared and developed. Since then, another historical event has taken place - it was finally conquered by neural networks and deep training. Among the tasks of processing natural language (Natural Language Processing, NLP) machine translation was one of the first to receive a strict statistical basis - back in the early 1990s. But in the field of in-depth training, he turned out to be a relatively late participant. In this post we are the team of Yandex for machine translation - we discuss why it took so long and what new possibilities ...
+ 0 -