# Syllabus: AINS6004 Natural Language Processing

## Catalog Description

Builds NLP systems spanning text processing, embeddings, transformers, RAG, conversation, and evaluation.

## Course Structure

Each week includes readings, a lecture/slide sequence, an executable lab, and an applied deliverable. Students maintain a reproducible project record and submit work through the LMS or GitHub workflow selected by the instructor.

## Weekly Schedule

| Week | Topic | Essential Question | Deliverable |
|------|-------|--------------------|-------------|
| 1 | Text preprocessing and linguistic signals | What is lost and gained when language becomes data? | Lab notebook + assignment brief |
| 2 | Embeddings and semantic similarity | How do vector representations support semantic operations? | Lab notebook + assignment brief |
| 3 | Language modeling foundations | What does a language model learn from context? | Lab notebook + assignment brief |
| 4 | Transformers for NLP tasks | How are transformer encoders and decoders adapted to applications? | Lab notebook + assignment brief |
| 5 | Retrieval-augmented generation | How does retrieval change reliability and accountability? | Lab notebook + assignment brief |
| 6 | Conversation design and tool use | How do dialogue systems coordinate context, tools, and user intent? | Lab notebook + assignment brief |
| 7 | Evaluation for NLP systems | Why are output quality and factuality hard to measure? | Lab notebook + assignment brief |
| 8 | NLP system deployment review | What makes a language system ready for use? | Lab notebook + assignment brief |

## Assessment

| Component | Weight |
|-----------|--------|
| Weekly labs and notebooks | 30% |
| Applied assignments | 35% |
| Participation and technical critique | 15% |
| Final synthesis portfolio | 20% |

## Graduate Expectations

Submissions must show technical reasoning, evidence awareness, clear limitations, and responsible use of AI assistance. Code and analysis should be reproducible enough for instructor review.