Multilingual Natural Language Processing — Overview

Course: Multilingual Natural Language Processing (5204MNLP6Y) — 2026/27 Sem. 1, Period 1 Programme: MSc Artificial Intelligence, UvA Credits: 6 EC Instructor: Christof Monz (Informatics Institute) Time zone: Europe/Amsterdam

Prerequisites

  • Basic math: linear algebra, calculus, probability
  • Basic machine learning: train/dev/test split, linear regression, classification
  • Basic implementation experience

Course Content

Provides an overview of NLP problems where multilinguality plays a role:

  • Multilingual scenarios: NER for multiple languages, language-independent parsing, cross-lingual classification
  • Crosslingual scenarios: Machine translation, crosslingual QA, crosslingual unlearning, crosslingual reasoning

This course is NOT: a comprehensive intro to ML or NLP — it focuses specifically on multilingual/crosslingual aspects.

Assessment

ComponentWeightDeadlineNotes
Mini ProjectTBDReport & code: 2026-10-09 12:00; Presentations: Week 7Teams of ~5; pick by 2026-09-04 17:00
ExamTBDTBD

Notes

  • Monz always requires pass on both assignments and the exam. Realistically, you’ll need at least a 5.5 on each.
  • The weekly schedule is structure-light; most of the work is project-driven with lectures as supplements.

Weekly Schedule

Week 1 — Overview & NLP Landscape

TopicLecturerReadingsNotes
L1Introduction to NLP, Applications, Course AdminMonzMNLP-L01 - Overview
Key DatesTeam formation & problem selection by Sep 4

Week 2 — Project Kick-off

TopicLecturerReadingsNotes
ProjectLiterature review, data inspection, baseline model selection

Week 3 — First Model

TopicLecturerReadingsNotes
ProjectImplement first model, evaluate, debug

Week 4 — Refinement

TopicLecturerReadingsNotes
ProjectRefine or try alternative model; dropout, layernorm, residual connections

Week 5 — Error Analysis

TopicLecturerReadingsNotes
ProjectSecond model refinement, error analysis, conclusions

Week 6 — Finalize

TopicLecturerReadingsNotes
ProjectLoose ends, 4-page report, presentation slides

Week 7 — Presentations

TopicLecturerReadingsNotes
ProjectTeam presentations (10-15 min + 5 min Q&A)

Key Deadlines

DateWhat
2026-09-04 17:00Submit team + problem choice (Google Sheet)
2026-10-09 12:00Submit report (4 pp PDF), slides, code (GitHub link)
Week 7Presentations

Mini Project — What Makes a Good One

  • Good succinct description of most relevant research papers
  • Good description of data preprocessing/settings
  • Good motivation of neural architecture choices
  • Thorough evaluation under different settings/architectures; error analysis
  • Report focusing on most relevant findings: what works and what doesn’t, why
  • MNLP - Mini Project — full project description, schedule, deliverables, and evaluation criteria

Resources

  • Mini Project slides (PDF in Assets/)
  • Canvas modules: Overview, Mini Project

  • Mini Project slides (PDF in Assets/)
  • Canvas modules: Overview, Mini Project