Sanyanggae LogoSanyanggae

AI

Starting with familiar features, we explore what AI does and how to read its results. Along the way, we connect data and mathematics, ideas inside models and uses in the real world.

Stories to read in order

  1. An illustration placing message cards being divided into two sorting trays beside a separate reply card and pencil
    An AI-generated explanatory illustration comparing message classification and reply generation. It is not an actual service screen, neural-network design or accuracy experiment result.

    6 min read

    How does AI pick out advertising messages?

    Follow a message to explore rules and machine learning, classification and reply generation, and ways to read AI results.

    Read the first story
Where will these stories lead?

Starting with the basics, we will gradually explore principles, research, and building and running things. These are future directions; only articles already published are available to read.

  • Where do predictions come from?

    Distinguish rules from examples, and building a model from using it.

  • The ingredients of a model

    Connect code, functions, vectors, matrices, probability and statistics when the story calls for them.

  • Language and images, choices and actions

    Continue into search and reasoning, prediction and neural networks, data splits and generalization.

  • Ideas in research papers

    Look at model assumptions and experimental results in language, vision, reinforcement learning and other fields.

  • AI in the real world

    Broaden the view to reproducing papers, testing hypotheses, data rights and quality, and model evaluation, deployment and operation.

Browse all computer and AI stories