The idalab seminars: our contribution to the AI discussion

For a while now, starting in March 2016, you have received information about the idalab seminars via our newsletter, Twitter, Linkedin or our Blog. For those who have never heard, or never attended – and always wondered – we want to take a brief moment to present to you what idalab seminars are all about, where we want to go with them and how we’d like you to be involved.

An open source project

idalab seminars are an “open source project”, in the sense that they are open to all interested parties, regardless if you have a background in Data Science and AI, or just want to have a drink with us and get to know us better. The ideas that are being presented are critically discussed, and the presentation material is being made publicly available on SlideShare.

We established the idalab seminars, because we believe that in order to maintain a self-critical and ever questioning attitude towards your work, it is important to engage in a close dialogue with scholars, business experts and thought leaders from your respective field, thereby gaining new perspectives and generating fresh insights.

What’s been there so far?

idalab seminars have been close to our hearts from the very beginning, but to start with, we only had capacities to arrange them on an irregular basis. However, ever since we moved into our new office at Potsdamer Straße 68, we have been delighted to be able to host a seminar once a month. We’re glad that the event has managed to develop a small crowd of regulars as well as continuingly attract new faces. Former seminars up until today include:

#1 – March 2016: Explaining Deep Neural Networks Classification Decisions – Grégoire Montavon

#2 – May 2016: Interpretation of multivariate machine learning models and time series connectivity analyses in the presence of correlated noise – Stefan Haufe

#3 – October 2016: From flow charts to neurons – structure and flexibility in dialogue systems – Alan Nichol

#4 – January 2017: Towards Scalable Fraud Detection – a journey from Sklearn to Spark – Stanimir Dragiev

#5 – August 2017: Canonical Trends – Detecting Trends in Web Data – Felix Bießmann

#6 – December 2017: ‘Lean’ Training Data: An Incremental Approach to Supervised Machine Learning – Jeremiah Lewis and Edouard Malet

#7 – January 2018: Next generation product search at Zalando using deep learning – Duncan Blythe

#8 – February 2018: Data Science meets art! – Predicting bestselling books – Ralf Winkler.

So far, idalab seminars presented academic insights and exceptional company showcases. This will perhaps always be the format’s main focus, but we’re open for there to be also seminars that ask questions about Data Science and AI that go beyond academia and business, attending to political implications, ethical dimensions, sociological effects and the influences of AI on art.

Where you come in

There are three ways in which we’d like you to be a part of the idalab seminars:

1. We organise everything, you are here to learn! You are always welcome to join the seminars or share word about the next one with your colleagues, friends and family. You can find out about upcoming seminars on our blog, Linkedin, Twitter and our Newsletter.

2. If you want to share your research, your work, your ideas – we provide the platform and you can seize the opportunity. Your talk should be about 30 minutes long and you should be prepared to answer plenty of curious questions. Seminar attendees consist of our contacts in business, academia and public institutions and always form a favourable, eagerly interested and highly diverse audience.

3. If you happen to have a super exciting topic and/or speaker in mind, please don’t hesitate to tell us about it. After all, as all open source projects, the idalab seminars will only really thrive with community involvement!

We’re looking forward to seeing you soon. For all questions regarding the idalab seminars, feel free to contact


Featured Image: A lecture on pneumatics at the Royal Institution, London. Coloured etching by J. Gillray, 1802. via Wellcome Collection.