PyData Warsaw #8: Deep & Machine Learning + After Party

This will be the 8th PyData Warsaw regular meetup at Centrum Szkoleniowe Adgar Ochota near Warszawa Zachodnia station (English: Warsaw West).

Location: http://en.adgarochota.pl/contact/location

room number: "Event Room"

Doors open at 18:00, talks start at 18:30pm, about 9pm we move to a pub. We are ready to host 150 folk in the room so there may be plenty of people to discuss data science questions with!

Please remember to unRSVP if you realize you can't make it - it will help a lot for our crew.

And make sure you follow @pydatawarsaw for any updates and early announcements.

First talk:
Andrzej Pacuk - Job recommendation challenge

The goal of RecSys Challenge 2016, was to recommend job offers for users of business oriented social network. I will describe the dataset, evaluation details and present our approach, which ranked 2nd.

Second talk:
Piotr Migdał - First steps in deep learning with Keras

Recents progress in deep learning, especially in image recognition, is fascinating. But... how to set our foot it this fast-paced word of possibilities?

Keras is a high-level deep learning library in Python, which can use both TensorFlow and Theano backends. It simplifies creation of artificial neural networks, making it good for learning, development and experimentation.

We can use Keras to create neural networks from scratch, or re-use pre-trained networks such as VGG16. Do to its simplicity, and popularity, it is easy to find readable examples. For the same reason when I teach practical deep learning - I start with Keras (as during this talk!).

Third Talk:
Rafał Hryciuk (CodiLime)- Neptune - Machine Learning Platform

The presentation will start with a brief description of a popular workflow for many data science projects. Next, common problems with running such projects will be described. The main part of the presentation will be a live demo of Neptune. Neptune is a machine learning platform that helps data scientists efficiently manage and monitor experiments and projects.

https://deepsense.io/neptune-machine-learning-platform/

Fourth Talk:
Dominik Choma - Building advanced stream processing pipelines with Apache Flink.

As you know data is most valuable when made available as soon as important events occur. We will show how to create low latency and high throughput analytic pipeline build with Apache Flink. Thanks to such solutions we can analyze data on the fly and create added value to act faster.

Agenda:

18:00 - 18:30 - Doors open

18:30 - 18:35 - Introduction

18:35 - 19:05 - First Talk

19:10 - 19:40 - Second Talk

19:40 - 19:50 - Break

19:50 - 20:20 - Third Talk

20:25 - 20:55 - Fourth Talk

20:55 - 21:00 "Speed Dating"

21:00 - 24:00 - After Party with our Partner, drinks for free !!!

PS1: default language is English, but there may be some exceptions from the rule

PS2: presentation part is mainly Python focused but not only, We expect to host a number of guests working with R, Scala and other languages.

direct contact: pszwed@gmail.com, mparfieniuk@mesoslabs.com

See you !