AI Summit Vienna – an afternoon of deep learning discussions
Finally caught a moment to reflect on Monday’s AI Summit Vienna. It was a wonderfully organised one-day rundown of global machine learning trends & local companies diving into the field, hosted at the WU (Wirtschaftsuniversität Wien) in a building perfectly suited for such futuristic talks.
The conference started with a high-level talk from Google’s Dave Elliott on the current use cases for AI in Google, TensorFlow as their attempt at standardising the tooling and some future goals, e.g. AutoML – learning to learn automatically (a bit of a mind-bender, this one).
Next up, Tomáš Mikolov, an AI research scientist over at Facebook and the creator of word2vec gave a more technical talk to the satisfaction of all the geeks in the crowd 🙂 One of the highlights was his fastText solution that dethrones the aforementioned word2vec method.
After the break, Sepp Hochreiter, a professor in Linz and the inventor of Long short-term memory (LSTM) recurrent neural networks (RNN) gave another higher-level talk on the capabilities of AI and motivated people to keep persisting, even if it takes years for their inventions to get recognised. He also taught us that Linz is about to become the next Mountain View 😉
Ulla Kruhse-Lehtonen, the CEO of DAIN Studios gave some great insights on introducing data science into a company – the importance of being smart and leveraging your data 🤓. She also gave some good tips to all aspiring data scientists.
Lastly, Zalando’s Calvin Seward brushed up our knowledge of the fundementals (backpropagation, GANs) in a very clear and funny manner. He also talked a bit about how they do data science at Zalando, locating clothing items in photos and the like.
There was also a poster showcase where 8 local Austrian startups doing AI pitched their work:
All in all, good talks and a great networking opportunity. All the talk slides are on the AI Summit Vienna webpage and you can watch the video recordings on YouTube.