Week 7 notes completed
This commit is contained in:
26
CM3060 Natural Language Processing/Week 7/Week 7 Notes.md
Normal file
26
CM3060 Natural Language Processing/Week 7/Week 7 Notes.md
Normal file
@ -0,0 +1,26 @@
|
||||
# Lexical semantics
|
||||
*The meaning of words*
|
||||
|
||||
## NLP Tasks
|
||||
* Classifications tasks (e.g. spam detection)
|
||||
* Sequence tasks (e.g. text generation)
|
||||
* **Meaning** tasks. We study meaning in this section
|
||||
|
||||
Useful for:
|
||||
* Information retrieval
|
||||
* Question answering
|
||||
* Topic modeling
|
||||
|
||||
What do words mean?
|
||||
* Look in a **dictionary**
|
||||
|
||||
## WordNet
|
||||
|
||||
A lexical database:
|
||||
* Nodes are **synsets** (sets of synonyms)
|
||||
* Correspond to **abstract concepts**
|
||||
* **Polyhierarchical** structure, looks like a tree, but is flexible, nodes can have more than one parent.
|
||||
|
||||
Using WordNet we can programmatically:
|
||||
* Identify **hyponyms** (child terms) and **hypernyms** (parent terms, we go up the tree for more general terms)
|
||||
* Measure **semantic similarity**
|
||||
Reference in New Issue
Block a user