AEVB
E835244
AEVB (Auto-Encoding Variational Bayes) is a foundational variational inference framework that combines neural networks and probabilistic modeling to learn latent representations of data, most notably underpinning variational autoencoders (VAEs).
All labels observed (1)
| Label | Occurrences |
|---|---|
| AEVB canonical | 1 |
How this entity was disambiguated
This entity first appeared as the object of triple T10023642 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: AEVB Context triple: [Auto-Encoding Variational Bayes, shortTitle, AEVB]
-
A.
EVB
EVB (Edge Virtual Bridging) is an IEEE networking standard that defines mechanisms for managing and integrating virtualized network interfaces on edge switches and servers in data center environments.
-
B.
AEV
AEV is the common abbreviation for Augsburger Panther, a professional ice hockey club based in Augsburg, Germany.
-
C.
VAB
VAB is the commonly used abbreviation for NASA’s Vehicle Assembly Building, the massive structure at Kennedy Space Center where rockets are assembled before launch.
-
D.
ÁVH
ÁVH was the secret police and state security organization of communist Hungary, notorious for its role in political repression and surveillance during the early Cold War era.
-
E.
AVE
AVE is Spain’s high-speed rail service, connecting major cities like Madrid and Barcelona with fast, long-distance trains.
- F. None of above. chosen
- G. Unsure - the case is ambiguous/there is not enough information to decide.
NED2
Entity disambiguation (via description)
gpt-5-mini-2025-08-07
Target entity: AEVB Target entity description: AEVB (Auto-Encoding Variational Bayes) is a foundational variational inference framework that combines neural networks and probabilistic modeling to learn latent representations of data, most notably underpinning variational autoencoders (VAEs).
-
A.
EVB
EVB (Edge Virtual Bridging) is an IEEE networking standard that defines mechanisms for managing and integrating virtualized network interfaces on edge switches and servers in data center environments.
-
B.
AEV
AEV is the common abbreviation for Augsburger Panther, a professional ice hockey club based in Augsburg, Germany.
-
C.
VAB
VAB is the commonly used abbreviation for NASA’s Vehicle Assembly Building, the massive structure at Kennedy Space Center where rockets are assembled before launch.
-
D.
ÁVH
ÁVH was the secret police and state security organization of communist Hungary, notorious for its role in political repression and surveillance during the early Cold War era.
-
E.
AVE
AVE is Spain’s high-speed rail service, connecting major cities like Madrid and Barcelona with fast, long-distance trains.
- F. None of above. chosen
Statements (46)
| Predicate | Object |
|---|---|
| instanceOf |
probabilistic machine learning method
ⓘ
variational inference framework ⓘ |
| abbreviationOf | Auto-Encoding Variational Bayes NERFINISHED ⓘ |
| application |
density estimation
ⓘ
generative modeling ⓘ missing data imputation ⓘ semi-supervised learning ⓘ unsupervised representation learning ⓘ |
| approximates | intractable posterior distributions ⓘ |
| arxivIdentifier | arXiv:1312.6114 ⓘ |
| assumes |
differentiable generative model
ⓘ
reparameterizable latent variables in standard form ⓘ |
| basedOn |
stochastic gradient variational inference
ⓘ
variational Bayes ⓘ |
| citationCountCategory | highly cited ⓘ |
| coreIdea |
optimize a variational lower bound using stochastic gradients
ⓘ
use an encoder network to amortize inference over latent variables ⓘ |
| field |
deep learning
ⓘ
machine learning ⓘ probabilistic modeling ⓘ |
| fullName | Auto-Encoding Variational Bayes NERFINISHED ⓘ |
| generativeNetworkType | decoder network ⓘ |
| inferenceNetworkType | encoder network ⓘ |
| influenced |
deep generative models research
ⓘ
representation learning research ⓘ |
| inspired | many VAE variants ⓘ |
| introducedBy |
Diederik P. Kingma
NERFINISHED
ⓘ
Max Welling NERFINISHED ⓘ |
| introducedInPaper | Auto-Encoding Variational Bayes NERFINISHED ⓘ |
| learns | latent representations of data ⓘ |
| notableContribution |
introduction of the reparameterization trick for VAEs
ⓘ
unification of neural networks and variational Bayes for latent variable models ⓘ |
| optimizes |
ELBO
ⓘ
evidence lower bound ⓘ |
| publicationYear | 2013 ⓘ |
| relatedTo |
Bayesian deep learning
NERFINISHED
ⓘ
amortized variational inference ⓘ |
| trainingObjective | maximization of ELBO ⓘ |
| typicalLikelihood | neural network likelihood model GENERATED ⓘ |
| typicalPrior | multivariate Gaussian prior GENERATED ⓘ |
| underpins |
VAE
NERFINISHED
ⓘ
variational autoencoder NERFINISHED ⓘ |
| uses |
latent variable models
ⓘ
neural networks ⓘ reparameterization trick ⓘ stochastic gradient descent ⓘ |
How these facts were elicited
The pipeline generated the facts above by prompting gpt-5.1 with this entity's name + description and the instruction below.
Instruction
You are a knowledge base construction expert. Given a subject entity and a description of it, return factual statements that you know for the subject as a JSON list of dictionaries(triples), where keys must be "subject", "predicate" and "object". The number of facts may be very high, between 25 to 50 or more, for very popular subjects. For less popular subjects, the number of facts can be very low, like 5 or 10. # Requirements - If you don't know the subject at all, return an empty list. - If the subject is not a named entity, return an empty list. - Include at least one triple where predicate is "instanceOf". - Do not get too wordy. - Separate several objects into multiple triples with one object.
Input
Subject: AEVB Description of subject: AEVB (Auto-Encoding Variational Bayes) is a foundational variational inference framework that combines neural networks and probabilistic modeling to learn latent representations of data, most notably underpinning variational autoencoders (VAEs).
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.