liveBook · Manning

Prior probability

As explained in Build AI Drug Discovery Pipelines

A prior distribution is a probability distribution specified before observing or encoding the data in a probabilistic model.

A prior distribution is specified before data are observed or encoded. In a variational autoencoder, it provides a shared target distribution for each molecule’s latent representation. section 10.3

Using the standard normal prior N(0, I), training penalizes differences between each encoder-produced distribution and this shared target through KL divergence. section 10.3

Sampling a latent molecule

After training, draw a latent vector z from N(0, I) and pass it to the decoder to generate a molecule from a region the decoder understands. section 10.3

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Explains the standard-normal choice, KL penalty, and sampling role, but does not develop general prior-selection theory.
Examples
Uses the standard normal N(0, I) and decoded latent vectors as concrete examples.
What you can do after
Apply a standard-normal prior to sample latent vectors for molecule generation with a variational autoencoder.
How it connects
Connects prior probability to variational autoencoders, latent variables, KL divergence, decoder sampling, and the evidence lower bound.

Also explained in

Compare how 12 books explain prior probability →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.