3rd Applied AI Meetup Innsbruck

Name: 3rd Applied AI Meetup Innsbruck
Start: 2019-11-21T18:00:00+01:00
End: 2019-11-21T21:00:00+01:00
Location: WEI SRAUM. Designforum Tirol

Hosted By

Sarah N. and Stefan E.

Details

// 3rd Applied AI Meetup Innsbruck //

We're hosting the next Applied AI meetup in Innsbruck - at the 21.11., the doors of WEI SRAUM will open another time for us. We organized three short talks from AI practitioners who will showcase their applications. You will have the chance to get hands-on experience and there will be time to network, discuss and ask questions. We will provide food and drinks. The pre-requirement for participating is to be registered at the Meetup group and event.

Agenda
6 pm // open door & showcase stations
6:45 pm // welcome

7 pm // talk 1: Data acquisition and demand prediction in the hotel industry by Martin Schwarz, Data Scientist at RateBoard

In recent years, the hospitality industry changed significantly towards digital technology. Acquisition of data, however, can be challenging, considering key factors of booking are room price, season, competitors, events, weather, availability of comparable rooms and much more. RateBoard offers hotel demand estimation and rate recommendations on a daily base. In this talk, Martin will give a short introduction to their field of operation.

7:10 pm // talk 2: Optimizing Training for Convolutional Neural Networks by Georg Westner, Chief Software Architect at Swarm Analytics

The problem in training neural networks is that a lot of expensive GPUs (graphics processing units) resources cannot be utilized properly for it. Georg talks about how different optimization techniques can enable more effective training in less time, utilizing multiple GPUs and machines.

7:20 pm // talk 3: How to Achieve State-of-the-Art in NLP with Transformer Language Model Finetuning by Alexander Rietzler, Lead Scientist at DeepOpinion.ai

Alexander is presenting the learnings and best practices from their BERT-ADA method published in their paper "Adapt or Get Left Behind: Domain Adaptation through BERT Language Model Finetuning for Aspect-Target Sentiment Classification" (https://arxiv.org/abs/1908.11860). As the lead author of the paper, he will provide insights into how the DeepOpinion team achieved state-of-the-art results for aspect-based sentiment analysis (ABSA) and was able to rank first at the benchmark leaderboard https://paperswithcode.com/sota/aspect-based-sentiment-analysis-on-semeval.