translate

Natural Language Processing

From text preprocessing to BERT — the tasks, models, and evaluation methods that define how machines understand human language.
Co-Created by Kiran Shirol and Claude
Topics Tokenization Embeddings Classification NER Transformers Evaluation
info Recommended prerequisites:  AI Fundamentals  and  Deep Learning Fundamentals  (or equivalent knowledge of neural networks and attention).
play_arrow Start Learning summarize Key Insights dictionary Glossary 10 Chapters · High-Level
Section 1

Text & Representation

Why language is hard for machines, how to preprocess text, and the journey from sparse to dense representations.
1
psychology
What Is NLP?
The field, its history from rules to neural nets, the core tasks taxonomy, and why human language is the hardest data type.
arrow_forward Learn
2
cleaning_services
Text Preprocessing
Cleaning, normalization, tokenization (word, subword, character), stemming vs lemmatization, and building a preprocessing pipeline.
arrow_forward Learn
3
scatter_plot
Representing Text
Bag-of-words, TF-IDF, Word2Vec, GloVe, and the revolution from sparse vectors to dense embeddings.
arrow_forward Learn
Section 2

Tasks & Models

The core NLP tasks — classification, sequence labeling, generation — and the models that solve them, from Naive Bayes to BERT.
4
label
Text Classification
Sentiment analysis, spam detection, topic classification. Naive Bayes, logistic regression, SVMs, and the classical ML pipeline for text.
arrow_forward Learn
5
data_array
Sequence Labeling
POS tagging, named entity recognition, chunking. HMMs, CRFs, BiLSTM-CRF, and the IOB tagging format.
arrow_forward Learn
6
auto_awesome
Language Models & Generation
N-grams, perplexity, RNN/LSTM language models, beam search, sampling, and the path to neural text generation.
arrow_forward Learn
7
bolt
The Transformer Revolution
Encoder vs decoder vs encoder-decoder. BERT, GPT, T5 — how one architecture changed everything in NLP.
arrow_forward Learn
Section 3

Modern NLP

Transfer learning, evaluation metrics, and the modern landscape from instruction tuning to multilingual models.
8
sync_alt
Transfer Learning & Fine-Tuning
Pre-train then fine-tune, feature extraction vs full fine-tuning, task-specific heads, and the Hugging Face ecosystem.
arrow_forward Learn
9
analytics
NLP Evaluation
Accuracy, F1, BLEU, ROUGE, METEOR, BERTScore, human evaluation — and why measuring NLP is harder than it looks.
arrow_forward Learn
10
explore
The Modern NLP Landscape
Instruction-tuned models, few-shot and zero-shot NLP, prompt-based approaches, multilingual models, and where the field is heading.
arrow_forward Learn
explore

Explore Related Courses

layers
Deep Learning Fundamentals
Neural nets, CNNs, RNNs & Transformers
token
How LLMs Work
Transformers, Attention & Generation
terminal
Prompt Engineering
Steering & Controlling LLMs
analytics
LLM Evaluation
Benchmarks, Metrics & Testing