Word Embeddings from Deep Space Nine using Machine Learning by Nick Evers , Gretel De Paepe

Sisko - Man + Woman = ? Word embeddings are a way to represent words in a numerical manner so they can be fed into machine learning models. In this presentation we will walk through some of the most common strategies, such as Bag of Words, Word2Vec, fastText, GloVe and Bert. Using the scripts of Star Trek Deep Space Nine, we will explore the differences between these techniques and how to use them in a production environment. Examples available @ https://github.com/GretelDePaepe/deep_space_9_devoxx GRETEL DE PAEPE Gretel De Paepe has over 20 years of experience in the world of data. She started as a data analyst, grew into a data scientist and eventually emerged herself into machine learning. She also functioned as a data science manager for a number of years. What makes her skills as an ML specialist unique, is that they were built on the solid foundation of years and years of solving real data challenges in the real world. She worked in both 'Silicon Valley' type of tech companies as well as large institutions. The job experiences closest to her heart were the time she spent in the UAE working for Abu Dhabi Investment Authority, building a data analytics team from scratch and her current job as as staff Data Scientist at Collibra. She also looks back with fondness on her team winning the AI hackathon in Madrid while at IBM, setting up her one-person consultancy shop in Dubai Knowledge Village and being named one of the inventors on a ML patent for data classification. NICK EVERS Nick is an AI Engineer at Collibra building classification and recommender systems to organize customer metadata assets. With over a decade of software engineering experience, Java, Python and Linux is what he speaks. AI/ML and data science is what draws his attention. ------------------------------------------------------------ INTRO * visuals & editing by @Mercator * music : Avocado by Ephixa