PyData Zurich: X-ray deep learning, Analytics Platforms, ML operationalisation

Nov 1, 2018 · Schlieren, Switzerland

A new edition of PyData Zurich! (7PM, 1st of November, at Zühlke Engineering - Schlieren)

In this meetup we will have the following exciting talks:

*******
*Joel Niklaus, Lukas Zbinden - From chest to hand X-rays: Transfer learning for skeletal age prediction

The transfer learning method in machine learning aims at applying and leveraging knowledge gained from solving one task to a different but related task. In our experiment we study and try to exploit transfer learning from chest X-rays to hand X-rays for skeletal age prediction. The deep learning model is first trained on the large chest X-rays dataset and then transferred for training on the much smaller pediatric hand X-rays dataset. Finally, the model is used for age prediction on unseen hand X-rays to compete with radiologists.

In this talk we will outline the deep learning experiments performed and conclude with insights obtained from this research area.

*******
* Hendrik Schöneberg - Data Analytics Platform in the Cloud

The rise of Data Analytics applications requires new architectures enabling the work of Data Scientists while being capable of processing large volumes and varieties of data. Building such a platform in the cloud comes with many challenges: There is a plethora of decisions to make, vendors to compare and design principles, services and products to choose from.

In this talk I will illustrate a data analytics platforms working on top of a data lake in the Google Cloud Platform, show the architectural foundation, the core concepts, and highlight important learnings.

The example is based on a marketing customer which leverages large-scale Apache Beam pipelines on Google Cloud Dataflow to crunch billions of records for their business intelligence use cases.

*******
* Joel Akeret - From PoC to Prod: Operationalization of machine learning

We see a wide adoption of Data Analytics and Machine Learning in various industry sectors. Many very promising model prototypes are being developed but only few manage to move from research-like projects into productive systems, causing frustration both on business and data scientist side. In my talk I will explore possible reasons for this and discuss why machine learning brings additional complexity to the common challenges of software engineering. I will show how we use MLFlow and Apache Airflow to address some of the difficulties in the machine learning lifecycle.

*******
and as always:
*Lightning talks - Your special 5 minutes
Every speaker gets 5 minutes to share interesting insights, projects or other fun stuff.

Event organizers
  • PyData Zurich

    Anybody that appreciates the PyData/NumFOCUS mission - Bringing the world of open source data science & scientific users together! - is greatly appreciated to join our meetup group and if you can and want, also help out in some way to make our meetups happen and great! We wish to have interesting talks in a friendly and informal setting with snacks/food and beer, so that people have a reason to stay and get to know each other. Also, we aim to be an accessible, community-driven meetup (and eventually c

    Recent Events
    More

Are you organizing PyData Zurich: X-ray deep learning, Analytics Platforms, ML operationalisation?

Claim the event and start manage its content.

I am the organizer
Social
Rating

based on 0 reviews