Report on the conference "AI & BigData Lab"

On April 12, the first Al & BigData Lab conference was held in Odessa, dedicated to artificial intelligence and big data. If the problem of artificial intelligence has been investigated for a long time, the concept of “big data” has appeared relatively recently and remains quite vague to this day. The main goal of the event was primarily to exchange experience among developers, to unite the community for more effective development of projects on this subject, as well as to create and promote big data solutions and artificial intelligence in Ukraine. We planned to lay the foundation for creating an AI-club in the Odessa region (as it was done in Kiev and Kharkov) to support and develop this area in the region. It is safe to say that we succeeded.

The event, held in two streams, was attended by more than 150 people from 9 cities of Ukraine. The conference was hospitably received by HUBOdessa. Reports were selected for every taste, people plied from one hall to another, choosing topics that they were interested in.

The following reports were presented at the conference:

• “Natural Language Processing in Practice” ( Vsevolod Demkin, Grammarly )
  We considered models and algorithms, basic tools (from Bayes theorem to deep neural networks) that are used in NLP. Unlike conventional software engineering, in NLP most of the time and effort is spent not on writing a program, but on experiments. This is a specific type of activity that requires its own approaches. As a result, it was considered how to collect all this into a single whole and get the finished product on the example of solving the problem from the real world using publicly available data.
• “Data mining and information search: problems, algorithms, solutions” ( Alexander Krakovetsky, DevRain Solutions )
  The report addressed the problems of information retrieval - duplicate information, a huge amount of information garbage, a large amount of time searching for “non-trivial” data. The report was not devoted to a specific Data Mining area or algorithm, but touched on several areas from the Text Mining area, clustering, NLP, SEO to solve a specific task - reducing the search time for necessary information, as well as the features of the service a la Readability, and how they can improve search algorithms.
• “MagnetoDB: Key / Value storage, BigData in OpenStack” ( Sergey Kovalev, Ilya Sviridov, Mirantis )
• “DeepLearning: Complex Data Analysis in Simple Words” ( Sergey Shelpuk, SoftServe )
  This report was devoted to explaining the principles of deeplearning networks in simple words, algorithmic techniques that make this process efficient, and technological tools that allow building networks in the fastest and cheapest way.
• “What is data visualization, or why specialists working with data should be slightly artists” ( Vladimir Timashov, DataArt )
  The speaker went through different stages of visualization and approaches to the visual presentation of information, as well as reviewed technical solutions that are available today for custom development (d3.js, Raphael, Three.js), and ready-made Enterprise products (Tableau, QlikView). In addition, the use of visualization in real projects - made in DataArt, and world leaders in this field is highlighted.
• “AI & BigData for travelers” ( Yuri Kuznetsov, Senturia )
  The author spoke about the practice of working with big data in the online service and the tasks of a modern travel agency. A review of approaches to the organization of the system of services, as well as the prospects of machine learning in the industry.
• “Structural Equation Modeling Modeling in the R Environment” ( Alexey Gaevsky ).
  Overview of structural equation modeling packages. Package lavaan: features, syntax. Practical application: building models, fine-tuning model parameters, processing missing values.
• “Entrepreneurship in BigData” ( Vitaliy Goncharuk, Augmented Pixels )
  Using the example of the Social Catcher project (Semantic technologies for social networks), the author described what e-commers expect from BigData and what are the business difficulties and prospects for the development of projects in BigData.
• "Tweaking performance on high-load projects" ( Dmitry Dumansky, Cogniance )
  Thanks to his customer, the author was able to present to the public the architecture of his advertising project, which has more than 85 million active users, an average load of 11,000 requests per second and more than 20 billion requests per month. The main questions addressed: What lessons have we learned? What code improvements have been made to improve performance? How was the request processing speed increased 10 times? Why I do not like Amazon services? How to build reporting with 25 GB of data per hour and why do we use Cascading?
• Master class "Analysis of large amounts of data in the cloud using Microsoft Azure HDInsight" ( Anton Boyko, Microsoft )
  During a remote online video report, the author told about an approach to data analysis called Map-Reduce, spoke about Hadoop and its components, and also what HDInsight is, led a number of demonstrations of the HDInsight service.

In the conclusion of the conference, a round table was held, the topics for discussion on which were outlined and proposed during the conference. Decuscia came out stormy and rich.
')
Thanks to all visitors - we were interested with you, and we hope that the conference turned out to be interesting and useful for you.

Special thanks to our gold sponsor Microsoft , DataArt and Sigma Ukraine sponsors, as well as to all our partners and information partners.

Presentations from the conference can be found here .

Video recordings of the reports will soon be available on our YouTube channel .

You can stay up to date with the news like this:

• Like us on Facebook: www.facebook.com/GeeksLab.co
• Join the group on Linkedin: www.linkedin.com/groups?gid=7430607
• Trickling on Twitter: twitter.com/geeks_labs
• We even have VK: vk.com/geeks_lab
• Add circles to G +: https://plus.google.com/101789932390194916614

---

Continue accepting applications for participation in WebCamp 2014

From July 21 to July 24, in the framework of Odessa Innovation Week 2014, the V professional conference of web developers WebCamp 2014 will be held. Each conference day will be devoted to a separate direction:

• July 21 - WebCamp: Developer Day - a day dedicated to the development of backend and frontend.
• July 22 - WebCamp: Designer Day & Internet Marketing Day - in parallel, independently of each other, there will be a day dedicated to design issues, as well as an Internet marketing day.
• July 23 - WebCamp: Mobile Day - mobile day: it will be a question of developing applications for various platforms: Android, iOS, Windows Phone and others.
• July 24 - WebCamp: HighLoad Day & Project Management Day - for the first time on WebCamp there will be a day dedicated to high loads and parallel to it a day on managing projects and teams.

To become a speaker or to give a master class / workshop , you must fill in the application form by June 2 at http://bit.ly/1l7zCNC .

We invite to cooperate partners and sponsors.

Odessa Innovation Week - a week of innovative technologies, which will be held from 21 to 27 July. During 7 days, the participants will be waited by a number of interesting conferences, workshops and master classes, in which experts in the field of web technologies, mobile development, high loads and project management will discuss various professional topics, entrepreneurs will demonstrate their projects and receive feedback from mentors and coaches , start-up entrepreneurs will have the opportunity to create their start-up in 48 hours under the strict guidance of domestic and foreign mentors, as well as make new acquaintances, find partners, and, of course, x Good rest on the Black Sea.