Welcome on the demo page of MoNoise; a modular normalization model. The demo on this page demonstrates a 'translation' of chat texts to more canonical language. The pretrained models are best suited for Twitter-like data, since they are trained on normalized Tweets. More information about how this demo works can be found in the paper and the code.
To try the model, simply choose the language from the dropdown menu, type your text in the input box on the left and press submit. Note that you can tune the aggressiveness of the model using the slider on the left (which will only update after another submit).
This system is developed as part of the Parsing Algorithms for Uncertain Input project, more information can be found on the Home Page. For any questions or bug reports, please contact Rob van der Goot (firstname.lastname@example.org).