ffsshhttiikk What I do
- Apply linguistic theory to computation
- Build language models with linguistic structure
- Implement syntactic and semantic analysis
- Create cross-linguistic NLP systems
- Handle multilingual text processing
When to use me
Use me when:
- Building linguistically-aware NLP
- Multilingual applications
- Grammar and syntax processing
- Semantic representation
Key Concepts
Linguistic Analysis Levels
┌─────────────────────────────────────────────┐
│ Discourse │
│ (context, coherence) │
├─────────────────────────────────────────────┤
│ Semantics │
│ (meaning, representation) │
├─────────────────────────────────────────────┤
│ Pragmatics │
│ (intent, context, speech acts) │
├─────────────────────────────────────────────┤
│ Syntax │
│ (grammar, structure rules) │
├─────────────────────────────────────────────┤
│ Morphology │
│ (word formation, inflections) │
├─────────────────────────────────────────────┤
│ Phonology/Orthography │
│ (sounds, written form) │
└─────────────────────────────────────────────┘
Syntax Parsing
import spacy
nlp = spacy.load("en_core_web_sm")
doc = nlp("The cat sat on the mat.")
# Part-of-speech tags
for token in doc:
print(token.text, token.pos_, token.tag_)
# Dependency parsing
for token in doc:
print(token.text, token.dep_, token.head.text)
# Named entities
for ent in doc.ents:
print(ent.text, ent.label_)
Semantic Representation
- Formal Semantics: Lambda calculus, FOL
- Frame Semantics: FrameNet
- Semantic Primitives: Primitive theory
- Distributional Semantics: Word embeddings
Linguistic Resources
- WordNet: Lexical database
- UD: Universal Dependencies
- Penn Treebank: Syntactic annotations
- BabelNet: Multilingual lexical knowledge