ConceptAI & ML
Latent variable model
Models that condition neural network outputs on unobserved random variables or seeds to represent broad distributions.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Marginalization1Mixture of Gaussians1Neural networks1Probabilistic model1Probabilistic modeling1Probability distribution1Video prediction1
Opens up
Concepts that name this one as a prerequisite.
Lecturers
3Each in their own terms, from the easiest level.
Sergey Levine
Phillip Isola
Models that describe observed data using unobserved variables that capture the underlying attributes of the data.Yann LeCun
A latent variable model uses an extra variable to allow a system to represent multiple possible predictions.
Lectures that assume it
1Connections
12The lectures that define it or assume it, and the concepts next to it.
Concepts7
- Marginalization
- Mixture of Gaussians
- Neural networks
- Probabilistic model
- Generative adversarial networks
- Mixture model
Show 1 moreShow fewer
Lectures5
- Generative Models: Representation Learning Meets Generative Modeling
- Variational Inference, Part 1 (CS 285, Lecture 18)
- The Deep Learning Revolution: The Sequel (Turing Award Lecture)
- Imitation Learning III: Inverse RL & Generative Adversarial Imitation Learning (GAIL)
- Latent Variable models and Subset Smoothing