85578 - SEMINARS (1) (LM) (G.B)

Conoscenze e abilità da conseguire

The seminars are designed to introduce students to some specialized aspects of the training. The seminars are related to the deepening of advanced topics, in particular taught by specialists or professionals of some disciplines of the learning activities. The seminars will deal with the themes connected with the areas of learning: computer science; literary, linguistic, historical/cultural and related to the arts in the digital context; transversal: economics, law and communication.

Contenuti

Machine Learning for the Arts & Humanities

 

 Learning outcomes

This seminar offers an overview of recent trends in Machine Learning (ML) applications in the Arts & Humanities. The seminar will consist of a few classes introducing foundational ML concepts and methods, while most of the remaining classes will focus on diverse areas of the Arts & Humanities and discuss recent advances in related ML applications. The students will be asked to prepare in advance of classes, by reading the proposed literature and engaging in discussion.

After completing this course, the student can:

• Understand basic Machine Learning concepts and methods.

• Find software libraries that can be used to develop ML applications.

• Understand recent trends in ML applied to the Arts & Humanities.

• Evaluate whether ML could be used in an Arts & Humanities task.

Course contents

Machine Learning is increasingly used in the context of Arts & Humanities research and GLAM applications (Galleries, Libraries, Archives, Museums). Examples range from text recognition and information extraction from historical sources to image search and analysis on artwork collections, from automatic 3D reconstructions of built heritage to the automatic detection of archeological sites from satellite or drone images.

The breakdown of the topics is as follows (per week):

1. Week 1: Introduction to Machine Learning. We discuss the course setup, the fundamentals of machine learning, the types of ML tasks, the key components of an ML workflow, some foundational mathematical concepts, and linear regression.

2. Week 2: Introduction to Machine Learning, continued. We discuss the worked-out examples of linear regression, linear classification, and the Multi-Layer Perceptron (MLP), with implementations. We also mention software libraries like Sklearn, Pytorch, Huggingface, Gradio, Weights & Biases.

3. Week 3: Image-based tasks. We discuss image classification, object detection, and image search tasks.

4. Week 4: Image-to-text tasks. We discuss automatic text recognition.

5. Week 5: Text-based tasks. We discuss topic modeling, document classification, and entity recognition.

Note that this list of topics is tentative and might still change slightly.

Testi/Bibliografia

The list of readings will be provided at the start of the course.

The following book serves as a reference for ML fundamentals:

• Zhang et al., Dive Into Deep Learning, MIT Press, 2023. https://d2l.ai/index.html .

The following articles are examples of the kind of applications we will discuss during seminars:

• Assael, Yannis, Thea Sommerschield, Brendan Shillingford, Mahyar Bordbar, John Pavlopoulos, Marita Chatzipanagiotou, Ion Androutsopoulos, Jonathan Prag, and Nando De Freitas. “Restoring and Attributing Ancient Texts Using Deep Neural Networks.” Nature 603, no. 7900 (March 10, 2022): 280–83.https://doi.org/10.1038/s41586-022-04448-z .

• Bundzel, Marek, Miroslav Jaščur, Milan Kováč, Tibor Lieskovský, Peter Sinčák, and Tomáš Tkáčik. “Semantic Segmentation of Airborne LiDAR Data in Maya Archaeology.” Remote Sensing 12, no. 22 (November 10, 2020): 3685.https://doi.org/10.3390/rs12223685 .

• Cetinic, Eva, Tomislav Lipic, and Sonja Grgic. “Fine-Tuning Convolutional Neural Networks for Fine Art Classification.” Expert Systems with Applications 114 (December 2018): 107–18.https://doi.org/10.1016/j.eswa.2018.07.026 .

• Lombardi, Francesco, and Simone Marinai. “Deep Learning for Historical Document Analysis and Recognition—A Survey.” Journal of Imaging 6, no. 10 (October 16, 2020): 110.https://doi.org/10.3390/jimaging6100110 .

• Luthra, Mrinalini, Konstantin Todorov, Charles Jeurgens, and Giovanni Colavizza. “Unsilencing Colonial Archives via Automated Entity Recognition.” Journal of Documentation, January 31, 2023.https://doi.org/10.1108/JD-02-2022-0038 .

• Wevers, Melvin, and Thomas Smits. “The Visual Digital Turn: Using Neural Networks to Study Historical Images.” Digital Scholarship in the Humanities, January 18, 2019.https://doi.org/10.1093/llc/fqy085 .

Metodi didattici

Recommended prior knowledge

While no prior knowledge is required, the students will benefit from having attended 1st-year courses such as ‘Computational Thinking’.

Teaching method and contact hours

Lectures and seminar discussions. Attending students are expected to come prepared to class.

Modalità di verifica e valutazione dell'apprendimento

Individual oral exam on all course contents (100%). The student may, optionally, work on a short essay further exploring an application area that we discussed in class or another one of their choice. The essay has to be sent to the lecturer at least 5 days in advance of the oral exam date. The essay will, in this case, contribute to 40% of the final grade, and the oral exam for the remaining 60%. Essay guidelines will be provided at the beginning of the seminar. Students are encouraged to do the essay since this will allow them to explore a topic of choice and lighten their oral examination.

The exam for non-attending students is the same (oral exam or essay + oral exam), but with extra readings.

Strumenti a supporto della didattica

Slides, live coding, demonstrations, readings, and seminar discussions.

Classes are held in a classroom equipped with personal computers connected to the Intranet and Internet.

Orario di ricevimento

Consulta il sito web di Giovanni Colavizza