Graphical Models for Machine Learning and Digital Communication

By Brendan J. Frey. 1998

Subfields: Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Probabilistic & Bayesian Methods

Citations in the corpus

Deep AutoRegressive Networks (2013, background): “Sampling from the encoder works in exactly the same way but in the opposite direction, sampling each hidden unit from q ( H A DARN without a stochastic hidden layer but with an autoregressive visible layer is a fully visible sigmoid belief network (FVSBN; Frey, 1998).”
Density estimation using Real NVP (2016, background): “The ordering of the dimensions, although often arbitrary, can be critical to the training of the model [66].”
RNADE: The real-valued neural autoregressive density-estimator (2013, neutral): “Perhaps surprisingly, but consistent with previous results on NADE [5] and by Frey [28], randomizing the order of the pixels made little difference to these results.”
NIPS 2016 Tutorial: Generative Adversarial Networks (2016, background): “Fully visible belief networks Fully visible belief networks (Frey et al., 1996; Frey, 1998) or FVBNs are models that use the chain rule of probability to decompose a probability distribution over an n-dimensional vector x into a product of one-dimensional probability distributions:”
A Deep and Tractable Density Estimator (2013, methodology): “For the binary data case, Frey (1998) considered the use of logistic regression conditional models, while Bengio & Bengio (2000) proposed a single layer neural network architecture, with a parameter sharing scheme different from the one in the NADE model (Larochelle & Murray, 2011).”
Reweighted Wake-Sleep (2014, methodology): “Deep AutoRegressive Network (DARN): An autoregressive sigmoid belief network layer (Frey, 1998; Bengio and Bengio, 2000; Gregor et al. , 2014) is similar to an SBN layer but with the output units x i not being independent of each other, given the layer’s input y .”
Taking on the curse of dimensionality in joint distributions using neural networks (2000, methodology): “The LARC model [ 3 ] described below is also of this type, but with first-order dependencies only.”
Bayesian Compressive Sensing Via Belief Propagation (2008, background): “Because entries of are restricted to , encoding only requires sums and differences of small subsets of coefficient values of .”
Transformation Autoregressive Networks (2018, background): “It may be written as a special case of LAM (5) with tied weights: We note also that LAM and NADE are both related to fully visible sigmoid belief networks (Frey, 1998; Neal, 1992).”
The Neural Autoregressive Distribution Estimator (2011, methodology): “Better Bayesian networks can be obtained by using log-linear logistic regressors for the conditionals, also known as fully visible sigmoid belief networks or logistic autoregressive Bayesian networks (Frey, 1998).”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, background): “PixelCNN [104] and fully visible belief networks (FVBNs) [105], [106].”
Information Theory, Inference, and Learning Algorithms (2004, methodology): “See (Wiberg, 1996; Frey, 1998; McEliece et al., 1998) for further discussion of the sum–product algorithm.”
Sparse-graph codes for quantum error correction (2003, methodology): “This iterative algorithm is explained in (MacKay and Neal 1996; MacKay 1999; MacKay 2003; Frey 1998).”
Bayesian Spiking Neurons I: Inference (2008, background): “We propose that the basic meaning of a spike is the occurrence of new, unpredictable probabilistic information and that propagation of spikes in cortical networks corresponds to propagation of beliefs in a corresponding Bayesian network (Frey, 1998; Jordan, 1974; Weiss & Freeman, 2001).”