DSAIT4090 · Natural Language Processing MSc DSAIT · TU Delft · 2026–2027 · Period 2

Course materials

A course on the modern aspects of Natural Language Processing. You will learn how today's NLP systems work end to end — from tokenization and language modelling to transformers and large language models, and how they are fine-tuned, retrieved from, aligned, served, and evaluated — and how to build, probe, and critically assess them.

Lectures are on Thursdays. Teaching team: J. Yang (responsible), A. Anand. Assessment: 20% individual assignments + 80% group project.

Schedule

Thursdays · morning 08:45 (CEG Lecture Hall E) · afternoon 13:45 (EEMCS Chip)
DateMorningAfternoonAssignments
12 NovIntro to NLP + LogisticsYangCEG Lecture Hall E——
19 NovLexical Analysis + Language ModellingAnandCEG Lecture Hall E—A1
26 NovText Classification + EmbeddingsAnandCEG Lecture Hall ESyntax and SemanticsYangEEMCS ChipA1 due
3 DecNeural LM & Contextual EmbeddingsYangEEMCS ChipDiscriminative TaskAnandCEG Lecture Hall EA2
10 DecGenerative ModellingAnandCEG Lecture Hall EText Generation TasksYangEEMCS ChipA2 due
17 DecRetrieval-based LMsYangCEG Lecture Hall EAlignmentguestEEMCS ChipA3
7 JanInterpretabilityAnandCEG Lecture Hall EAgentsguestEEMCS ChipA3 due
14 JanLLM ServingAnandCEG Lecture Hall ERobustnessYangEEMCS Chipfinal project due
AnandYang/ guest

Assignments

Individual · auto-graded + short reflection · laptop-runnable (no GPU required)
A1

Tokenization & statistical LM

BPE, a bigram (add-k) LM, perplexity per-token vs per-byte, compared to a TULIP-served LLM (Gemma 4).

Group project briefs

Course-long group project (research & development) · 80% of the grade

Data & starter

Starter kits are on Brightspace; large models run on the TULIP pilot

Content note: some datasets (e.g. toxicity comments, Enron e-mails) contain offensive or sensitive material — handle with care.

Explainer videos

Short narrated walkthroughs · example lecture material
Language ModellingNarrated explainer · what a language model is and how it assigns probabilities to text.
TokenizationNarrated explainer · how raw text is split into tokens, and why BPE is used.