Lesson 4

Advanced Techniques

Two Improvements

  • Performance
    • Accuracy
    • Specificity
    • MCC
  • Scalability
    • More data
    • Faster results

Performance Improvements

- Word Order
- Grammar
- Semantics
- Smarter Dimension Reduction
- Learned Metrics
- Generalization (Clustering)

Tools

- gensim
    - approximate, incremental learning
- Dirichlet Allocation
- Word2Vec
- NLTK POS tagging
- Parsey McParseface
- neural nets
    - CNNs
    - Skip-grams
    - LTSTM

Generative Models

- Chat bots
  - Tay, 
- Assistants
  - Siri, Google Now