Speaker

Abstract

Nowadays, we have a variety of computational methods to process large text corpora. For example, we can discover topics in discourse, analyze the sentiment of social media discussions on relevant topics, or automatically detect hate speech and toxic language. Currently, many analyses and applications are implemented using so-called neural language models. These are deep neural networks that learn complex manifold relationships in languages based on large digital text collections like Wikipedia or online news collections. In this talk we will give an insight into how neural language models work and discuss how they can be used for interesting downstream tasks.

Type to search or press esc to cancel.