Question & Answering with Google's BERT Language Model
Below you will find our Q&A demo featuring BERT, one of the most talked about transformer language models (LM) released by Google. To date, BERT continues to be a performance leader on various NLP tasks researched by AI engineers worldwide. The advantages of a language model lies in its ability to transfer the knowledge it learned via pre-training down to custom and smaller datasets.
Our demo encompasses a Closed Domain Question and Answering (CDQA) model, which allows the user to ask questions about a selected piece of text. We've included example paragraphs with preset questions for you to try. However, you may also choose to ask your own custom questions! Also, don't forget you can also paste custom text to see how BERT performs on data it’s never seen. Just remember to ask questions about information that is explicitly discussed in text. Have fun, and follow us on Twitter for AI news and upcoming demos: @Quantum_Stat.