Final Review

• Final Topics
• Learning Objectives
  • Python/Programming
    • The re Module
    • Built-in Functions
  • NLP/CL
    • String Manipulation
    • N-grams
    • Part-of-Speech Tags
    • Machine Learning and Classification
  • Software Engineering
    • Regular Expressions
    • Advanced Functions

Final Topics

• Python/Programming
  • re module (re.match(), re.search(), re.fullmatch(), re.sub())
  • Built-in Functions (map(), filter())
• NLP/CL
  • Stemming and Lemmatization
  • Segmentation and Tokenization
  • N-grams and Collocations
  • Part-of-Speech Tags and Tagging Methods
  • Statistical Inference
  • Automatic Evaluation
  • Backoff Methods
  • Baseline Systems
  • Machine Learning
  • Supervised vs Unsupervised Learning
  • Linguistic Features for Machine Learning
  • Classification
  • Decision Trees
  • Entropy
  • Ethics
• Software Engineering
  • Regular Expressions
  • Higher-order Functions
  • Recursive Functions

Learning Objectives

Python/Programming

This category is for programming concepts, techniques, and structures as used in Python.

The re Module

re.search()

re.match()

re.fullmatch()

re.sub()

Built-in Functions

map()

filter()

NLP/CL

This category is for concepts and techniques related to natural language processing or computational linguistics.

String Manipulation

Stemming and Lemmatization

• What is the difference?
• When would you use one or the other?

Segmentation and Tokenization

• What is the difference (i.e., when is segmentation not tokenization)

N-grams

N-grams

• What is an n-gram?
• Why do we pad sequences when computing n-grams?
• How do you get n-grams of a sequence (using NLTK or just Python)?

Collocations

• What are collocations?
• What are they good for?

N-gram Language Models

• What can you do with an n-gram language model?
• What is the formula for a bigram model? (at least, in general terms)

Part-of-Speech Tags

Part-of-Speech Tags

• What are tag sets? Why are there different ones?
• Do all languages use the same tag set?

Tagging Methods

Machine Learning and Classification

Statistical Inference

• What are the basic requirements for statistical inference?
• Create a basic model for predicting, e.g., gender, part-of-speech, next words, etc.

Automatic Evaluation

• Why do we use automatic evaluation?
• How would you evaluate, e.g., a part-of-speech tagger?

Backoff Methods

• What makes a good backoff method?
• Why do we use backoff methods?

Baseline Systems

• What makes a good baseline system?
• Why do we use baseline systems?

Machine Learning

Supervised vs Unsupervised Learning

• What is necessary for supervised classification?

Linguistic Features for Machine Learning

• Give examples of linguistic features for tasks like POS-tagging, genre classification, etc.

Classification

• What is the classification task?
• What are some examples?

Decision Trees

• In general terms, how is a decision tree created?
• How is it applied to classify some new instance?

Entropy

• What is the formula for entropy?
• What does it measure?

Ethics

• Can you describe and give examples of the following kinds of problems?
  • exclusion
  • overgeneralization
  • overexposure / underexposure
  • dual-use

Software Engineering

This category is for practices of software engineering as well as programming concepts that are relevant to programming languages beyond Python.

Regular Expressions

Advanced Functions

Higher-order Functions

Recursive Functions