AI engineering glossary · Generative media multimodal
What is ELBO (evidence lower bound)?
Evidence lower bound (also called ELBO): The objective maximised when training a VAE: the expected reconstruction quality minus a KL divergence penalty pulling the encoder's distribution toward a prior. It is a tractable lower bound on the log probability of the data.
Evidence lower bound: an example
ELBO = E[log p(x|z)] - KL(q(z|x) || p(z)); a rising ELBO means better reconstructions with a well-behaved latent space.
Related terms
- Variational autoencoder
- Posterior collapse
- Beta-VAE
More Generative media multimodal terms
- FastSpeech
- Flamingo
- Flow matching
- Forward diffusion process
- Fréchet Inception Distance
- Generative adversarial network
- Grapheme-to-phoneme
- Guidance scale
Where you learn it
Evidence lower bound is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.