Equivalence of Generative and Log-Linear Models

48 citations recorded by the source corpus.

How researchers see this paper

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

Rapid adaptation for deep neural networks through multi-task learning (2015, methodology): “It is shown in [31], that a log-linear model is equivalent to a Gaussian model.”
Optimization Algorithms and Applications for Speech and Language Processing (2013, methodology): “The popular model functions that were constructed in Section II all qualify as -functions, given appropriate assumptions on the smoothness and concavity of the objective and on the properties of the sets and .”
Adaptation of context-dependent deep neural networks for automatic speech recognition (2012, methodology): “A substantial amount of work has been conducted to improve the performance of the GMM-HMM based ASR systems under the mismatched conditions.”
Learning small-size DNN with output-distribution-based criteria (2014, methodology): “The mapping from a log-linear model to a Gaussian model 5( ; 63, Σ3) is [15] Σ3 = − 12 (239 + Δλ9)? and 63 = Σ323 (8)”
Machine Learning Paradigms for Speech Recognition: An Overview (2013, background): “Finally, we note that with appropriate parameterization some classes of generative and discriminative models can be made mathematically equivalent [135].”
Investigations on an EM-Style Optimization Algorithm for Discriminative Training of HMMs (2013, methodology): “2 illustrates this inequality for two points .”