Tighter PAC-Bayes Bounds

By Amiran Ambroladze, Emilio Parrado-Hernández, John Shawe-Taylor. 2006

127 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Statistical Learning Theory

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 Statistical Learning Theory

Citations in the corpus

PAC-Bayesian Theory Meets Bayesian Inference (2016, positive): “That is, the provided results express upper bounds on E f ∼ ˆ ρ L D ( f ) for any (bounded, sub-Gaussian, or sub-gamma) losses.”
PAC-bayes bounds with data dependent priors (2012, background): “The encouraging results of Ambroladze et al. (2007), motivate a further use of the prior PACBayes bound.”
Chromatic PAC-Bayes Bounds for Non-IID Data (2009, background): “Their possible tightness, as witnessed by the work of Ambroladze et al. (2007), make them a possible route to perform model selection. They can also be viewed as theoretical tools to motivate and build new learning procedures. Unfortunately, up to now, PAC-Bayes bounds only hold for the case of clas”
Tighter PAC-Bayes bounds through distribution-dependent priors (2013, methodology): “We should finally distinguish between distribution defined priors and using part of the data to learn a prior and the rest to learn the function (Ambroladze et al., 2006).”
Robust Forward Algorithms via PAC-Bayes and Laplace Distributions (2014, methodology): “Another approach [1] is to split the training set into two parts, using the first to learn data-dependent prior, and the second to learn the posterior.”
A New PAC-Bayesian Perspective on Domain Adaptation (2015, unknown): “The taken approach is the one privileged in numerous PACBayesian works (e.g., Langford & Shawe-Taylor, 2002; Ambroladze et al., 2006; McAllester & Keshet, 2011; Parrado-Hernández et al., 2012; Germain et al., 2009; 2013), as it makes the risk of the linear classifier hw and the risk of a (properly p”
Information-Theoretic Generalization Bounds for SGLD via Data-Dependent Estimates (2019, methodology): “The idea of using data-dependent priors to obtain tighter bounds is standard in the PAC-Bayes literature [2, 10, 22, 27], but its utility in the present work is brought through by our introduction of data-dependent estimates.”
PAC-Bayesian Collective Stability (2014, methodology): “PAC-Bayesian analysis was introduced by McAllester (1999) and later refined by a number of authors (e.g., Langford and Shawe-Taylor, 2002; Seeger, 2002; Ambroladze et al., 2006; Germain et al., 2009).”
PAC-Bayesian Analysis of Co-clustering and Beyond (2010, methodology): “…of decision trees (Mansour and McAllester, 2000), Support Vector Machines (SVMs) (Langford and Shawe-Taylor, 2002; McAllester, 2003; Langford,2005; Ambroladze et al., 2007; Crammer et al., 2009; Germain et al., 2009), transductive learning (Derbeko et al., 2004), structured prediction (Bartlett…”
PAC-Bayesian learning of linear classifiers (2009, methodology): “1The idea of using a fraction of the training data to construct a prior has been proposed in (Ambroladze et al., 2006) for the problem of choosing the hyperparameter values of the SVM.”
Stability and Generalization in Structured Prediction (2016, methodology): “PAC-Bayesian analysis is often accredited to McAllester (1998, 1999), and has been refined by a number of authors (e.g., Herbrich and Graepel, 2001; Langford and Shawe-Taylor, 2002; Seeger, 2002; Ambroladze et al., 2006; Catoni, 2007; Germain et al., 2009; Lever et al., 2010; Seldin et al., 2012).”