← All teaching

2022–2023 · Courses

Machine Learning Approaches in Preservation Masterclass

Graduate masterclass on machine learning workflows for preservation and facade deterioration analysis.

Institution
Columbia University, GSAPP
Level
A6921-1
Period
2022–2023

About the course

This masterclass was planned and proposed for graduate students at Columbia GSAPP. It focused on introducing machine learning in a preservation context and culminated in a final project on identifying deterioration in the facades of heritage buildings.

I developed the syllabus and the course materials and structured the class so that students could connect technical ML workflows with preservation-specific questions.

Students learn to

  1. 01 Understand core ML workflows in a preservation context.
  2. 02 Develop applied pipelines for facade deterioration recognition.
  3. 03 Translate ML concepts into heritage-focused final projects.

How it runs

  1. Module 1 ML Foundations Introductory concepts and workflow design for graduate students.
  2. Module 2 Preservation Data Working with visual and building-related heritage datasets.
  3. Module 3 Project Development Developing a final project around deterioration detection in facades.

Assignments

  • Dataset Framing Define a preservation-oriented ML problem and prepare source data.
  • Final Project Build a project aimed at recognising deterioration in heritage building facades.
Next course Learning from Design Heritage: Investigation of Data-driven Methods