CS408

Natural Language Processing

Teaching machines to work with human language: representing text as vectors, language models, and the transformer architecture behind modern AI.

10 modules · 50 lessons · Practice after every lesson

Take first

Syllabus

  1. Module 1

    Language Data and Preprocessing

    • Language as Structured Data
    • Tokenization and Segmentation
    • Normalization and Morphology
    • Corpora, Annotation, and Data Quality
    • Evaluation Splits and Leakage

  2. Module 2

    Statistical Language Models

    • N-Gram Models
    • Smoothing
    • Perplexity
    • Class-Based and Cache Models
    • Limits of Count-Based Models

  3. Module 3

    Vector Semantics

    • Distributional Semantics
    • Count-Based Word Vectors
    • Predictive Embeddings
    • Similarity, Analogy, and Bias
    • Subword and Contextual Representations

  4. Module 4

    Sequence Labeling

    • Part-of-Speech Tagging
    • Hidden Markov Models
    • Conditional Random Fields
    • Named-Entity Recognition
    • Neural Sequence Tagging

  5. Module 5

    Syntax and Parsing

    • Constituency Grammars
    • Probabilistic Context-Free Grammars
    • Dependency Grammar
    • Transition and Graph-Based Parsing
    • Parsing Evaluation

  6. Module 6

    Meaning and Information Extraction

    • Lexical Semantics and Word Sense
    • Semantic Role Labeling
    • Coreference Resolution
    • Relation and Event Extraction
    • Knowledge-Base Construction

  7. Module 7

    Neural Sequence Models

    • Recurrent Neural Networks
    • Sequence-to-Sequence Models
    • Attention
    • Transformer Architecture
    • Position and Context Representation

  8. Module 8

    Pretrained Language Models

    • Pretraining Objectives
    • Tokenizers and Model Inputs
    • Fine-Tuning and Prompting
    • Instruction Tuning and Preference Optimization
    • Efficient Adaptation and Distillation

  9. Module 9

    NLP Applications and Retrieval

    • Machine Translation
    • Question Answering
    • Summarization
    • Information Retrieval and Dense Search
    • Retrieval-Augmented Generation

  10. Module 10

    Evaluation, Safety, and Linguistic Diversity

    • Automatic and Human Evaluation
    • Hallucination and Factuality
    • Bias and Representational Harm
    • Multilingual and Low-Resource NLP
    • Privacy, Copyright, and Responsible Deployment

Start Natural Language Processing.

No setup, nothing to install. Try the first lessons before you sign up.

Start Learning
Try first