Skip to content

Variational Autoencoder (VAE)

Definition
AI-generated

A variational autoencoder (VAE) is a generative autoencoder whose bottleneck encodes parameters of an approximate posterior over latents; training maximizes an evidence lower bound (ELBO) combining reconstruction fidelity and a penalty (typically KL to a prior), enabling sampling and smooth interpolation in latent space.

Topics
LLM and Agents Machine learning concepts
Related
Variational Inference (VI) Autoencoder Decoder

Why it matters in GWAS

VAEs model single-cell expression manifolds, simulate synthetic molecular profiles, and support perturbation prediction pipelines; combined with genetic covariates they can reveal structure but also absorb confounding. Claims of causal latent factors require designs that separate genetics, environment, and batch.

Example usage

"The generative baseline was a VAE fit to control cells before in silico knockout predictions were benchmarked against CRISPR screens."

Variational Inference (VI) Autoencoder Decoder Deep Generative Model Encoder In Silico Perturbation Prediction Neural Network Single-Cell Foundation Model
Related-term network

Click any node to open that term page.

Node order. The current term is first. Next come every page linked from Related terms, in the same top-to-bottom order as that list. Then new terms appear in two-hop layers: for each of those direct neighbors (in that same order), their related terms are appended if not already listed. A third pass does the same for two-hop nodes. The walk stops when a fixed maximum number of nodes is reached.

References

  • Kingma DP, Welling M. (2014). Auto-encoding variational Bayes. ICLR.
← Variation graph Variational Inference (VI) →

Last updated (UTC · Git history)