AI in Motion
Read the Lecture Hall ↗

Generative AIBeginner1:12 video5 chapters

Generative vs Discriminative Models — lecture notes

One kind of model learns where the boundary is; the other learns what the data looks like — and can create new examples.

▶ Watch the animated lecture

0:001. Introduction

Introduction — Generative vs Discriminative Models

Generative AI can write, draw and compose. What makes a model generative? It comes down to what the model learns about the data.

0:102. Two ways to learn

Two ways to learn — Generative vs Discriminative Models

A discriminative model only learns the boundary between classes: is this pink or blue? A generative model learns what each class looks like, here as a cloud of probability. Because it knows the shape of the data, it can draw brand new samples from it, the amber circles.

0:313. The difference

The difference — Generative vs Discriminative Models

Discriminative models learn the probability of a label given the data. Generative models learn the probability of the data itself. That is harder, but it lets them create new images, text and sound.

0:454. Families

Families — Generative vs Discriminative Models

Generative AI grew through several families: variational autoencoders in 2013, GANs in 2014, generative pre-trained transformers from 2018, diffusion models from 2020, and in 2022 Stable Diffusion and ChatGPT brought them to everyone.

1:005. Recap

Recap — Generative vs Discriminative Models

To recap. Discriminative models learn boundaries. Generative models learn the data itself, which lets them sample new examples. VAEs, GANs, diffusion models and language models are all generative.

Key takeaways

  • Discriminative models learn P(label | data); generative models learn P(data).
  • Generative models can sample brand-new examples.
  • Major families: VAEs, GANs, autoregressive models (LLMs) and diffusion models.
  • Generative modelling is harder but far more flexible.

Check yourself

  1. What does a discriminative model learn?
    Show answer

    The boundary between classes, P(label | data) — It focuses on separating classes.

  2. Why can generative models create new examples?
    Show answer

    They model the distribution of the data itself — Sampling from the learned distribution produces new data.

  3. Which is a generative model?
    Show answer

    A diffusion model — Diffusion models generate data.

Go deeper

© 2026 Janin A Apurba, CSE, AUST · Advanced ICT Officer, CNRS-UNHCR. All rights reserved. Notes for the animated lecture at https://ai-in-motion.vercel.app/watch/generative-vs-discriminative-models.html