We started with the concept of a full-day meetup in 2016. Our aim was to provide the entire community with the opportunity to participate at a sort of data science conference at the expense of a ticket to the movies. The fact that this actually came to life is only due to our great sponsors that pay for food, drinks, and location. We are tremendously humbled by the increasing demand for our event with 100 people in the first year, 200 in the second, and 400 in the third season.
Learn more about the hottest event about data and AI.
Unsupervised verb sense induction using Bayesian nonparametric clustering (Pólya Urn model) with Gibbs sampling. Applied to billion web pages, automatically discovered 437 senses of "enter" from syntactic contexts—no labeled data required.
Abstract etc. Due to insufficient audio quality it was not possible to render a transcript, hence no summary available.
Deep learning for industrial defect detection: ConvNets learn directly from raw pixels, beating handcrafted algorithms. Key insight—traditional ML saturates with data while deep learning scales continuously, essential for big data era. Width beats depth for simple defects.
Ethics in data analytics beyond legal compliance: Through case studies (Target pregnancy prediction, stalker apps, network monitoring), argues transparency and saying "no" to customers matters more than hiding behind consent and terms & conditions.
Abstract etc. Youtube failed to create a transcript due to the insufficient audio quality, no summary possible.