For this month's meetup and we are very excited to have Ioannis Kourouklides to talk about TensorFlow and Deep Learning, and Michalis Kitis to talk about Streaming Data Analysis using Apache Spark.

#Talk 1/2

Title:
An Introduction to TensorFlow for Deep Learning, Predictive Analytics, and Data Science

Speaker:
Ioannis Kourouklides

Abstract:
TensorFlow recently became the “de facto” software library for Deep Learning and it is arguably the most popular one. It is open source and it was released in 2015 by Google to make it easier for Developers to design, build, and train Deep Learning models in Python.

Everyone is welcome at this talk and the presentation will start from the very basics and will then be adopted according to the preferences of the audience. Beginners and advanced coders are equally encouraged to attend, as the presentation will be very flexible and it will try to cater all needs. All Data Professionals, Academics and Enthusiasts are highly encouraged to attend, as TensorFlow has several aspects behind it, for both the Production side and the Research side. A brief introduction about what is Machine Learning and about what is Deep Learning, will also be presented in the beginning of the presentation.

Apart from TensorFlow, other libraries which work either with or on top of TensorFlow, such as Keras, will also be discussed. Similar to previous meetups, this will be a fully hands-on tutorial, with practical examples. Join us at this meetup to learn how to train and use an advanced classifier in a just a few lines of Python code simply by using TensorFlow.

Bio:
Ioannis Kourouklides pursued his four-year Master of Engineering (MEng) at Imperial College London in the fields of Electronic and Information Engineering, with specialization in Αrtificial Intelligence, Optimization, Signal and Image Processing. He graduated in 2014 with a First Class Honours and his thesis involved applications of Computer Vision and Machine Learning in Mobile Robotics. PhD Candidate at Data Science Laboratory.

website: www.kourouklides.com

#Talk 2/2

Title:
Streaming Data Analysis using Apache Spark for breaking news on twitter

Speaker:
Michalis Kitis

Abstract:
The purpose of the talk is to demonstrate the potentials of data mining in social networks. The social network that has been used is Twitter for which two separate implementations took place.

• Analyse and Visualise static data for the Greek elections of January 2015
• Analyse streaming data from twitter using the technology Apache Spark.

The first implementation constitutes a system that collects static data through Twitter API analyzes and visualize them. The datasets that were collected consists of twitter follower’s ids for political parties. This implementation is a classic example of how data mining has applied
over the last years. However new challenges took place like data mining on big data or analyze streaming data in real time which need a different approach.

Analyzing big data streaming was not a work that can be done a few years ago. However new technologies like Apache Spark make this possible. In the second implementation, we present a streaming data analysis for real-time data occurred by twitter. In order to achieve our goal, we used the new promising technology Apache Spark which is a framework for Apache Hadoop.

Bio:
Michalis Kitis graduated in 2016 from the Department of Computer Engineering and Informatics at University of Patras. He worked for a year as a Computer Engineer and now he is a freelance web developer. He is interested in Data Science and web technologies.

SPONSORS/ SUPPPORTERS

Cyprus University of Technology / cut.ac.cy