CS-595-052

Machine Learning and Statistical Natural Language Processing

Class resources

On the article summary page, click on either "PS" (for Postscript) or "PDF" (for PDF) to get the article itself.

• Lecture 1:
  Tutorial: N-gram models and smoothing
  Advanced n-gram models
• Lecture 2:
  Tutorial on HMMs
  British National Corpus tagset
  Building a tagged corpus
  An HMM POS tagger
• Lecture 3:
  Transformation-based POS tagging
• Lecture 4: Word-sense disambiguation
  Using a thesaurus and probabilistic modeling
  About senses per discourse
  
• Shallow parsing:
  Linguistic argument for shallow parsing
  Transformational shallow parsing
  Memory-based shallow parsing
  MBSL shallow parsing
  More memory-based
  
• PP Attachment:
  Winnow-based
  Memory-based
  Rule-based
  
• Information Extraction:
  IE overview
  Learning coreference in Japanese
  Learning coreference in English
  IE in German
  Using IE in text categorization
  FASTUS
  

Slides (in Powerpoint)

1. Machine Learning, Statistical Natural Language Processing, N-gram Models
2. HMM Part-of-speech tagging, Viterbi algorithm
3. EM algorithm, Transformational learning
4. Word sense disambiguation
5. Clustering and unsupervised word sense disambiguation
6. Shallow parsing
7. PP Attachment
8. Information Extraction

1. Assignment 2
   POS tagged corpus
2. Assignment 3
3. Assignment 4