Reading like a Machine

• An interactive workshop

  This interactive workshop uses visual content, grids, and buttons in a step-by-step Google Colab Notebook which combines textual explanations and executable Python code to introduce humanities students to key concepts in machine learning. The activities require no previous knowledge of programming, and utilize clickable graphics to enable learners to train and evaluate models capable of distinguishing between handwritten digits. By the end of the workshop, learners should have a working knowledge transferrable to more complex artificial intelligence tools.

• 

  Reading like a machine: Understanding machine learning through visual experiments

  A hands-on introduction to concepts like features, weights, and biases using scikit-learn and the MNIST dataset. Students learn the mechanics of training, prediction, and evaluation through visual interaction.

  Launch in Colab → View on GitHub

About this Workshop

• Learning objectives

  In this introductory workshop I created during my postdoctoral fellowship at Princeton, students train a simple classifier to recognize handwritten digits using a prewritten Python notebook. Through interactive widgets, they manipulate parameters and observe how changes in input affect model performance. This exercise opens space for larger discussions about what it means for a computer to “learn” and how models process human writing. By the end, students not only understand the inner workings of a basic machine-learning pipeline but also approach AI systems with new conceptual clarity—learning key machine learning concepts like features and weights, and how to apply them to more complex situations.

• For a more detailed technical explanation, please see the GitHub README file .