This meetup will showcase the wide range of NLP activities at Axel Springer.
General-purpose word embeddings with StarSpace
We'll present StarSpace (https://github.com/facebookresearch/StarSpace), a model introduced last year by Facebook Research for training general-purpose embeddings. After an overview of some of the applications presented by the Facebook team in their paper (https://arxiv.org/abs/1709.03856) with special reference to document similarity evaluation, we will present our own use case in the context of content-based recommendation.
Speaker: Richard Lawrence, Senior Data Scientist at Axel Springer
NLP applications @ Welt.de
In this talk, we will speak about what problems we are solving with the help of NLP for one of the most popular newspapers in Germany - welt.de. We will present the different use cases, our approaches and the specific techniques, algorithms and libraries we are using. While taking an applied perspective, we will also touch technical topics such as conditional random fields, lemmatization vs. stemming or how are editors participate in training and optimizing our models.
Speaker: Stoyan Stoyanov, Senior Software Engineer at WeltN24 GmbH