We invite you to Moscow Data Science Meetup on September 1
September 1, we are pleased to invite you to the next meeting of the Moscow Data Science community, where you can exchange practical experience in solving data analysis problems and chat with like-minded people. For one day the program is very rich, there will be two sections of reports, and among speakers there will be two speakers from neighboring countries. Also, the guests of the mitap will have an excursion around the office of Mail.Ru Group and a drawing of souvenirs. Join now! The meeting program under the cut.
“Psychotyping users of social networks”
Mikhail Firulik, Mail.Ru Group
The psychological type is a structure, a framework of a person’s personality. The psychotype forms the worldview of a person, lies at the heart of his behavior, including affecting the perception of information. In Russia, in various industries, the three main concepts of personality typing have received the most practical application: Big5, MBTI, and Socionics. The report will consider the features of the definition of these typologies and their modeling.
“Stochastic Computing Graphs in NLP”
Maxim Kretov, MIPT
Consider the formalism of stochastic computational graphs (graphs that contain sampling from distributions at nodes). To calculate the gradient of the loss function in such graphs, the usual back-propagation algorithm for the error is no longer suitable, and more complex methods need to be applied. As an example, there are various ways to train a simple seq2seq model.
“TensorFlow Sequence Prediction: The Engineering Side of the Question”
Denis Dus, InData Labs
We will discuss the features of building TensorFlow graphs for the effective training of architectures that work with sequences. We will try to answer the basic questions that an engineer may have in the process of building a model: what format for storing data to choose, how to organize the process of receipt of training and validation data in a model, how to work more efficiently with sequences of different lengths, how to save a model for future use, and others.
“Kaggle's Amazon from Space: Classification of Satellite Imagery”
Arthur Cousin, Avito
In the report, we will consider the methods and tricks of training deep convolutional neural networks using the example of the Planet: Understanding the Amazon from Space solution to the kaggle competition. In this competition, the ODS community teams Russian Bears and ods.ai took 2nd and 7th places, respectively, out of more than 900 participants. We will also analyze well-known top solutions.
“Multiple Time Series Forecasting”
Vitaliy Radchenko, Ciklum
Let's talk about what features exist in the choice of metrics, validation, data set generation and feature generation for the task of predicting multiple time series. Let us analyze the “groups” of signs and approaches to modeling using several real cases as examples, as well as problems that need to be avoided, and what should be paid attention first of all.
Gathering of participants and registration: 18:00
Beginning of reports: 18:30
Address: office of Mail.Ru Group, Leningradsky Prospekt 39, p. 79.
To participate, you must register . For those who will not be able to attend in person, a video broadcast will be organized: section 1 , section 2 .