Description

Machine learning has gained a lot of traction in recent years. It can solve problems previously believed impossible to tackle. It is no longer a privilege of the few to utilize these technological advancements. During my talk, I will explain how you can utilize these services without a Statistics Ph.D. By the end of the talk, you will know what Google Cloud offers to you, whether you are a product manager, an analyst or a computer engineer.

Description

Have you ever wondered how a big data stack look like and what are the building blocks? In this talk, I will try to guide you through how Prezi’s petabyte scale data platform looks like and showing what worked for us to manage this enormous amount data and building out a devops like culture with the data pipelines to make sure our small data infra team is not drowning in maintenance and support. 

Prezi

Prezi is the presentation platform that helps you connect more powerfully with your audience and customers. Unlike slides, Prezi's single, interactive canvas encourages conversation and collaboration, making your overall presentation more engaging, persuasive, and memorable. Prezi's latest offering, Prezi Next, is a full-lifecycle presentation platform so everyone can easily create visually stunning presentations, deliver them in a more natural and conversational way, and analyze their effectiveness. Founded in 2009, and with offices in San Francisco, Budapest, and Riga, Prezi now fosters a community of over 100 million users and Prezi presentations have been viewed over 3.5 billion times. Its investors include Accel Partners, Spectrum Equity, and TED conferences. For more information, please visit www.prezi.com.

Description

Cancer is a complex and diverse disease arising from gene mutations that can happen in a myriad of ways, therefore, finding one single cure is an impossible task. Designing effective, targeted therapies for given cancer types requires an understanding of the mutation patterns that gave rise to the disease. Using machine learning techniques and cell simulations, it is now possible to catch a glimpse of those patterns, which, paired with biological insight, have a vast potential for facilitating cancer drug development. That's the big picture, but what's the first step? When does machine learning come into play, and what can you do with it? This talk is going to pull back the curtains on a real-life research project aiming to uncover the genetic patterns affecting drug resistance in patients with chronic lymphocytic leukemia - from idea to implementation and beyond.

Turbine AI

Turbine enhances cancer drug discovery processes by taking trial and error out of the laboratory and moving it onto servers that can run millions of simulated experiments in minutes.

Description

The future is automated, everyone knows that. How to bring cutting edge research project into a robust and most importantly safe Level 4 autonomous vehicle is a question only a few people know the answer to. In our next meetup, 3 presenters will give you an insight into blockage detection of environmental sensors, construction of a „data lake” and testing & validation of AI-based algorithms all recent developments along the way to the top of „maturity ladder”.

1.How Do Self-Driving Cars Learn?

Pogácsás Sándor - Artificial Intelligence Engineer

How can we teach cars to drive safely?

What is the role of data and machine learning pipelines in this challenge?

In this talk, we will explore some of the tools and open datasets used by autonomous driving research teams.

Can self-driving cars have more experience than a professional human driver? Let us find out!

2. Not enough data…

Dr. Vizi Zsolt - Machine Learning Engineer

What shall we do, when we need special data, but gathering and processing them is too expensive? 

How can we multiply our measurement data?

This talk aims to introduce to data generation method for developing perception in self-driving cars. However, the following question  has to be clarified first: does a synthetically improved training data lead to a good result?