PAC-Bayes Risk Bounds for Stochastic Averages and Majority Votes of Sample-Compressed Classifiers

By François Laviolette, Mario Marchand. 2007

34 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering

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

1. Christoph H. Lampert
2. Anastasia Pentina
3. François Laviolette
4. Pascal Germain
5. Jean-Francis Roy
6. Alexandre Lacasse
7. Mario Marchand
8. Pierre Alquier

In Control and Systems Engineering

1. Christoph H. Lampert
2. Anastasia Pentina
3. François Laviolette
4. Jean-Francis Roy
5. Pascal Germain
6. Alexandre Lacasse
7. Mario Marchand
8. Ben London

Citations in the corpus

• A PAC-Bayesian bound for Lifelong Learning (2013, methodology): “For each task we use 2% of the data (approximately 20 images) available for collie class and the same amount of images from the another task, such that data between different tasks does not overlap.”
• Lifelong Learning with Non-i.i.d. Tasks (2015, methodology): “However for 0/1-loss the error of a Gibbs classifier is bounded from below by half the error of the corresponding majority vote predictor [18, 19] and therefore twice the right hand side of (25) provides a bound for deterministic classifiers.”
• Risk bounds for the majority vote: from a PAC-Bayesian analysis to a learning algorithm (2015, methodology): “This second approach is based on the sample compression theory (Floyd and Warmuth, 1995; Laviolette and Marchand, 2007; Germain et al., 2011).”