Fundamentals of AI and Machine Learning
for Byzantine Studies I
Postgraduate Programme — Level 7
LEARNING OUTCOMES
Upon completion, students will be able to:
1)
Explain the main families of AI methods, including rule-based systems, machine learning, deep learning, and generative AI.
2)
Recognize the main types of data used in Byzantine Studies and describe why historical, textual, visual, spatial, and material evidence often resist simple computational treatment.
3)
Understand basic machine learning concepts such as training data, features, labels, models, prediction, classification, overfitting, and generalization.
4)
Interpret basic evaluation concepts such as accuracy, precision, recall, F1 score, false positives, and false negatives without needing to implement them mathematically.
5)
Identify common sources of uncertainty in Byzantine research data and explain how uncertainty should be documented.
6)
Critically assess AI outputs by asking questions about training data, assumptions, confidence, bias, provenance, and failure modes.
7)
Describe the role of digitisation, OCR, HTR, metadata, and research data management in AI-assisted Byzantine Studies.
8)
Design a small, realistic AI-assisted research plan for a Byzantine topic, including question, data source, potential method, risks, and verification strategy.
COURSE SYLLABUS
13 Modules
Week 01 | What is AI? A humanities-oriented introduction
AI as a family of methods, not a single technology. Rule-based systems, machine learning, deep learning, and generative AI explained through examples from Byzantine Studies.
Week 02 | Byzantine evidence as data
Texts, manuscripts, inscriptions, seals, coins, images, objects, monuments, places, and people as research data. Structured vs unstructured data; born-digital vs digitised material.
Week 03 | Uncertainty, ambiguity, and provenance
Disputed dates, uncertain attributions, variant names, damaged sources, incomplete corpora, and contested identities. Why uncertainty is a scholarly feature, not merely a technical problem.
Week 04 | Machine learning concepts
Training data, features, labels, models, classification, clustering, prediction, generalisation, and overfitting explained through humanities analogies.
Week 05 | How to evaluate AI outputs
Accuracy, precision, recall, F1, false positives, and false negatives. How to read simple evaluation tables and why high scores do not automatically mean scholarly reliability.
Week 06 | Data preparation and documentation
Cleaning, normalization, metadata, identifiers, data dictionaries, licensing, and provenance. Introduction to good research data habits for humanities projects.
Week 07 | Digitization foundations: OCR, HTR, and transcription
Conceptual introduction to OCR and HTR. Why historical scripts, abbreviations, polytonic Greek, damaged manuscripts, and editorial conventions create difficulties.
Week 08 | AI outputs and critical reading
A practical checklist for evaluating OCR transcriptions, entity lists, image labels, summaries, and classifications. Confidence, bias, hallucination, and domain mismatch.
Week 09 | Generative AI and scholarly responsibility
What large language models can and cannot do. Acceptable support uses, verification workflows, citation checking, and transparent declaration of AI assistance.
Week 10 | AI-assisted research design
From research question to data audit, method choice, pilot test, interpretation, and communication. When computation is useful and when it is not.
Week 11 | Survey of specialized methods in the programme
Overview of NLP, GIS, computer vision, knowledge graphs, mixed reality, and interactive applications.
Week 12 | Mini-project workshop
Students develop a modest AI-assisted research plan: question, data source, tool/method, verification strategy, risks, and expected scholarly contribution.
Week 13 | Presentations and synthesis
Presentation of mini-project plans and collective reflection on what AI can responsibly contribute to Byzantine Studies.
ASSESSMENT
Student Evaluation
40%
Writtent Assignments
40%
Final Project
20%
Workload — ECTS Distribution
250 Hours Total
Lectures
39
Guided exercises / workshops
30
40
55
40
26
20
Course Total
250
Recommended Bibliography
Suggested bibliography:
- Mitchell, M. Artificial Intelligence: A Guide for Thinking Humans. Farrar, Straus and Giroux, 2019.
- Russell, S. Human Compatible: Artificial Intelligence and the Problem of Control. Viking, 2019.
- Kelleher, J. D. Deep Learning. MIT Press, 2019.
- Karsdorp, F., Kestemont, M., and Riddell, A. Humanities Data Analysis: Case Studies with Python. Princeton University Press, 2021.
- Flanders, J., and Jannidis, F. (eds.). The Shape of Data in Digital Humanities. Routledge, 2019.
- Baca, M. (ed.). Introduction to Metadata. Getty Research Institute, selected chapters.
- The Programming Historian. Selected beginner-friendly lessons on data literacy, OCR, text analysis, and digital humanities methods.
- European Commission. Ethics Guidelines for Trustworthy AI, 2019.
- UNESCO. Recommendation on the Ethics of Artificial Intelligence, 2021.
Related academic journals:
- International Journal of Digital Humanities (Springer)
- Journal on Computing and Cultural Heritage (ACM)
- Digital Scholarship in the Humanities (Oxford University Press)
- Journal of Open Humanities Data
- Journal of Cultural Heritage