Data Science meetup at Avito office June 24

On June 24, we gather Data Science specialists in our Moscow office in order to exchange experience in creating reference services. At the meeting, we will summarize the Avito competition held at the Dataring.ru site to build a recommendation system for announcements: reward the winners and ask them to tell you more about their decisions. In addition, the program includes interesting reports from representatives of Yandex. Dzen, OZON.ru and, of course, Avito. Details under the cut!

Reports

At the beginning of the meeting, we will hear speakers who will share their experience in creating advisory services.
')

Machine learning in Zen

The first speakers will be speakers from Yandex: Evgeny Sokolov and Dmitry Ushanov. Evgeny is the head of the quality group for recommendations and content analysis of Yandex. Dzen. Dmitry is a senior developer of the Zen service, and before that he was engaged in the development of the linguistic components of the search: search engine witches and object answers.

Yandex.Den is a personal recommendation service that aggregates news and media content from a large number of sites, and also allows authors to publish directly to the platform. At all stages of building recommendations, from collecting content and filtering to ranking, machine learning is used. In recommender algorithms, two main types of signals are used: user feedback and semantic proximity of content. The speakers will analyze several unusual examples of taking these signals into account: how to use models with hidden variables for texts in matrix expansions and how to properly form factors based on them; how to account for user clicks using sports rating systems; how to combine explicit and implicit user feedback.

Recommendations in OZON.ru

Leading analyst at OZON.ru Ksenia Boksha, without going too hard in mathematics, will talk about the various types of recommendations that the marketplace provides to its users: related products, accessories, bundles, personal recommendations, recommended categories of products for queries, recommended search queries, recommendations in the basket. In addition, from the report you will learn about the technology stack that the team uses, as well as about the nearest plans.

What tasks solves the team recommendations in Avito

At one of the past meetings, the head of the recommendations unit Vasiliy Leksin already told that under the hood, the recommendations in Avito. This time, together with analyst Mikhail Kamenshchikov, they will talk about what tasks the unit of recommendations solves, why there was a competition to build an advisory model, what results they expected to get at the end of the competition, what worked out and what didn't. In addition, they will share their experience of participating in the RecSys Challenge 2017 .

Competition

Next, we will proceed to summarizing the competition , award the winners and ask them to tell some details about their approaches and decisions.

A few words about the contest itself. In this dating, participants were asked to build their recommendation system for ads based on an activity history of about 600,000 users within 6 days. After receiving the training sample, participants had to predict events of user interaction with the ad, which could be of 4 types: click on the ad, send a message to the seller, add the ad to favorites and contact the seller. Correct predictions of events of different types had different weight.
The final results of the competition have already been published on the portal DataRing.ru , and the decisions of the winners are validated.

check in

In order to get to the meeting, you need to register and get confirmation . Please register with your real name and do not forget to take your passport or driver's license with you on the day of the event.

We start the reports at 12:30 and plan to complete the meeting by 16:00. During the break, you can have a snack wok.

Address: Moscow, Lesnaya, 7, 15th floor (BC "White Gardens"). Entrance from Forest Street. The nearest metro station is Belorusskaya Koltsevaya.

See you!