The increasing availability and throughput of information flowing through social media outlets or internet-based news poses challenges for users and the society as a whole. Shared content is quite often not fact-checked and erroneous, leading to users assigning equal value to both factually correct and incorrect pieces of information. The implications of this phenomenon are severe.
In this workshop, we will take action against misinformation and explore ways in which an average user can evaluate the reliability of the content. Through this workshop, we will gain an understanding of various machine learning-related concepts. Expect diving with us into natural language processing, classification/clustering and neural networks. After we cover the necessary theoretical background, we will work hands-on with the Fake News Challenge dataset and apply these methods in practice, trying to find the best model. We’ll use visualizations to interpret these methods along the way.
As a bonus, we will spend the last few minutes of the workshop on providing you with an overview of common tools used to defend users against misinformation.
Participants will be given an introduction to natural language processing and some major machine learning concepts such as typical tasks (classification/clustering) and methods (neural networks). They will obtain hands-on experience with Python based NLP-libraries and pipelines and learn how to solve classification problems with the help of neural networks in the context of the fake news challenge task.