i'm a student and i'm new to NLP.

I want to build an Automated Scoring system which is in Indonesian Language using BERT. The system is expected to be able to measure the similarity of an answer(e.g: student answer) with a reference answer and give the value of similarity.

But i'm still not clear about how the right step-to-step process to build the system using BERT. Since my system is going to process texts that are in Indonesian and i'm going to build and test the system on Google Colab, my main questions are :

  1. For the preprocessing, is it necessary for using a library such as NLTK or is it enough just to use BERT's default Tokenizer?
  2. Should i build my own vocabulary and my own sentence vectors? (For later, i will use this google colab tutorial to build my own sentence vectors if it's really necessary to build them on my own)
  3. Will my system work properly if i only use the BERT-Base Multilingual Uncased as the language model?

And maybe there's any complete article/tutorial about this? I appreciate any helpful answers. Thank you so much in advance.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.