Working with Google BERT: Elements of BERT

Expected Duration
Lesson Objectives
Course Number
Expertise Level


Adopting the foundational techniques of natural language processing (NLP), together with the Bidirectional Encoder Representations from Transformers (BERT) technique developed by Google, allows developers to integrate NLP pipelines into their projects efficiently and without the need for large-scale data collection and processing. In this course, you'll explore the concepts and techniques that pave the foundation for working with Google BERT. You'll start by examining various aspects of NLP techniques useful in developing advanced NLP pipelines, namely, those related to supervised and unsupervised learning, language models, transfer learning, and transformer models. You'll then identify how BERT relates to NLP, its architecture and variants, and some real-world applications of this technique. Finally, you'll work with BERT and both Amazon review and Twitter datasets to develop sentiment predictors and create classifiers.

Expected Duration (hours)

Lesson Objectives

Working with Google BERT: Elements of BERT

  • discover the key concepts covered in this course
  • compare approaches to supervised and unsupervised learning in NLP
  • define the concept of language models and recognize their purpose
  • list multiple legacy language models and their use cases
  • describe how deep learning neural networks can create language models
  • name state-of-the-art language models and recognize their utility
  • describe the purpose of language representation in NLP pipelines and neural network models
  • outline how developers make use of transfer learning
  • describe the concept and purpose of transformer models
  • describe Google BERT and how it is used in NLP products
  • outline Google BERT's architecture and list use cases of its variants
  • name multiple real-world problems in NLP that are solved by Google BERT
  • work with an Amazon review dataset and Google BERT to develop sentiment predictors
  • work with a Twitter dataset and Google BERT to create disaster Tweet classifiers
  • summarize the key concepts covered in this course
  • Course Number:

    Expertise Level