Discovering the NLP built-in algorithms in Amazon SageMaker
SageMaker includes four NLP algorithms, enabling supervised and unsupervised learning scenarios. In this section, you'll learn about these algorithms, what kind of problems they solve, and what their training scenarios are:
- BlazingText builds text classification models (supervised learning)or computes word vectors (unsupervised learning). BlazingText is an Amazon-invented algorithm.
- Latent Dirichlet Allocation (LDA) builds unsupervised learning models that group a collection of text documents into topics. This technique is called topic modeling.
- Neural Topic Model (NTM) is another topic modeling algorithm based on neural networks, and it gives you more insight into how topics are built.
- Sequence-to-sequence (seq2seq) builds deep learning models predicting a sequence of output tokens from a sequence of input tokens.