Parameter learning with truncated message-passing

By Justin Domke. 2011

53 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, 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 Artificial Intelligence

In Computer Vision and Pattern Recognition

In Probabilistic & Bayesian Methods

Citations in the corpus

Structured Attention Networks (2017, methodology): “The differentiability of inference algorithms over graphical models has previously been noted by various researchers (Li & Eisner, 2009; Domke, 2011; Stoyanov et al., 2011; Stoyanov & Eisner, 2012; Gormley et al., 2015), primarily outside the area of deep learning.”
Gated Graph Sequence Neural Networks (2015, background): “An analogy can be drawn between our adaptation from GNNs to GG-NNs, to the work of Domke (2011) and Stoyanov et al. (2011) in the structured prediction setting.”
Iterative Neural Autoregressive Distribution Estimator NADE-k (2014, background): “This learning-to-cope was introduced recently in 2011 by Stoyanov et al. (2011) and Domke (2011).”
Deep Unfolding: Model-Based Inspiration of Novel Deep Architectures (2014, methodology): “In [12, 13], tree-reweighted belief propagation (TRW-BP) and mean ﬁeld inference were unrolled and trained via gradient descent.”
Algorithm Unrolling: Interpretable, Efficient Deep Learning for Signal and Image Processing (2019, methodology): “For concrete examples, Domke studies and compares gradient descent and heavyball and limited-memory Broyden–Fletcher–Goldfarb–Shanno algorithms for image labeling and denoising applications.”
Blending Learning and Inference in Conditional Random Fields (2016, background): “Other approaches for blending CRFs appear are developed by Domke (2011).”
Learning to Pass Expectation Propagation Messages (2013, background): “Perhaps the most superficially similar line of work to the approach we present here is that of inference machines and truncated belief propagation [2, 3, 10, 9], where inference is done via an algorithm that is structurally similar to belief propagation, but where some parameters of the updates are…”
Generic Methods for Optimization-Based Modeling (2012, methodology): “Other work similar in spirit includes fitting graphical model parameters in terms of the marginals produced after a few message passing iterations (Domke, 2011, Stoyanov et al., 2011), fitting a sparse coding model in terms of a few iterations of coordinate descent (Gregor and LeCun, 2010), and…”
Minimum-Risk Training of Approximate CRF-Based NLP Systems (2012, methodology): “For a common pixel-labeling benchmark in computer vision, Domke (2011) shows that this procedure improves training time by orders of magnitude, and slightly improves accuracy if the same number of message-passing iterations is used at test time.”
Training energy-based models for time-series imputation (2013, methodology): “This model is similar to the Recurrent Temporal Restricted Boltzmann Machine (Sutskever et al., 2008) but in our model the units that define the energy are in a separate layer and the visible variables are not independent given the hidden variables.”
New Era for Robust Speech Recognition (2017, methodology): “Another important statistical beamformer is the multichannel Wiener filter (MWF) [7].”
Learning Graphical Model Parameters with Approximate Marginal Inference (2013, background): “marginals has little practical impact, but this can still be enough to prevent learning from succeeding [51].”
Human Pose Estimation with Fields of Parts (2014, methodology): “The FoP model is built upon advances from three separate domains: efficient inference for segmentation [18], parameter estimation with approximate inference [7,8], and expressive PS models [34].”
Mutual CRF-GNN for Few-shot Learning (2021, background): “Conditional Random Field (CRF) [27, 11, 12, 10, 6] is a popular probabilistic model to infer various dependencies within a image group or pixels in a image.”
Approximation-Aware Dependency Parsing by Belief Propagation (2015, methodology): “Smith and Eisner (2008) previously showed how to run BP in this setting (by calling the inside-outside algorithm as a subroutine).”
Randomized Optimum Models for Structured Prediction (2012, background): “Domke (2011) deﬁnes probability models around a ﬁxed number of iterations of belief propagation.”