A fun little app I am trying to put together to make it easy to utilize/test the model I am working on as a part of my Neural Ad Hominem research project.
My hope is to eventually have an app that pulls fresh tweets from twitter, classifies them, and points out examples of ad hominem on the website. Then it will hopefully allow viewers to flag classifications as correct or incorrect and then send the tweet back and add it as part of the training data. From there we should re-train the model every so often and re-load the model into the app when it is done training.