Project Sherlock

Artificial Intelligence

Generative Models

Models that produce samples rather than labels.

10 topics · 16 curated works

Topics

Reading in Generative Models

16

A way in

  1. Start here

    No prior grounding assumed.

    MIT 6.S191 (2023): Deep Generative Modeling

    Alexander Amini (MIT 6.S191) · 2023

    Introduces autoencoders, VAEs, GANs and diffusion as different answers to the same question — how to learn a model that can generate new, realistic…

    +1 more at this level

  2. Then

    Assumes you know the vocabulary.

    Hierarchical Neural Story Generation

    Fan, Lewis & Dauphin · 2018

    Introduces top-k sampling and a hierarchical generate-the-premise-then-the-story pipeline, arguing generic beam search decoding is the wrong tool for…

    +3 more at this level

  3. Go deeper

    Primary sources and full treatments.

    Auto-Encoding Variational Bayes

    Diederik P. Kingma & Max Welling · 2013

    Moving the randomness outside the network makes the sampling step differentiable, so a latent-variable model can be fitted by ordinary gradient…

    +9 more at this level

12 of 16 works

Elsewhere in Artificial Intelligence