.png){kind=icon}
Return
Simpler PAC-Bayesian bounds for hostile data
.png){kind=icon}
OA
AI
DOI：10.1007/s10994-017-5690-0.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
PAC-Bayesian learning bounds are of the utmost interest to the learning community. Their role is to connect the generalization ability of an aggregation distribution to its empirical risk and to its Kullback-Leibler divergence with respect to some prior distribution . Unfortunately, most of the available bounds typically rely on heavy assumptions such as boundedness and independence of the observations. This paper aims at relaxing these constraints and provides PAC-Bayesian learning bounds that hold for dependent, heavy-tailed observations (hereafter referred to as hostile data). In these bounds the Kullack-Leibler divergence is replaced with a general version of Csiszar's f-divergence. We prove a general PAC-Bayesian bound, and show how to use it in various hostile settings.
Keywords:
PAC-Bayesian theory
Dependent and unbounded data
Oracle inequalities
f-divergence
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2.9
Papers:
2.7K
Citations:
3.4W
Organization
Cited Papers
ShareSaveShareSaveMode of action of the gramicidin S analogs lacking hydrophilic amino acid residues on biomembranes.OAAI
OAAIShareSaveOptimization of Texture-Modified Yellowfin Sole (Pleuronectes aspera) by Enzymatic Treatment and Superheated Steam Treating to Improve Quality Characteristics.png)
ProcessesOAAI
ProcessesIF0
OAAIShareSaveEquilibria in solutions of amines and fatty acids with relevance to the corrosion inhibition of iron
PREAI
ShareSaveShareSaveAdaptive demodulation synchro-extracting transform for bearing time-varying fault feature extraction
PREAI
ShareSave
View more