Benign overfitting in multiclass classification: All roads lead to interpolation

K Wang, V Muthukumar… - Advances in Neural …, 2021 - proceedings.neurips.cc
The growing literature on" benign overfitting" in overparameterized models has been mostly
restricted to regression or binary classification settings; however, most success stories of …

Risk bounds for the majority vote: From a PAC-Bayesian analysis to a learning algorithm

P Germain, A Lacasse, F Laviolette… - arXiv preprint arXiv …, 2015 - arxiv.org
We propose an extensive analysis of the behavior of majority votes in binary classification. In
particular, we introduce a risk bound for majority votes, called the C-bound, that takes into …

[图书][B] Model selection and error estimation in a nutshell

L Oneto - 2020 - Springer
How can we select the best performing data-driven model? How can we rigorously estimate
its generalization error? Statistical Learning Theory (SLT) answers these questions by …

Model selection and error estimation without the agonizing pain

L Oneto - Wiley Interdisciplinary Reviews: Data Mining and …, 2018 - Wiley Online Library
How can we select the best performing data‐driven model? How can we rigorously estimate
its generalization error? Statistical learning theory (SLT) answers these questions by …

On the proliferation of support vectors in high dimensions

D Hsu, V Muthukumar, J Xu - International Conference on …, 2021 - proceedings.mlr.press
The support vector machine (SVM) is a well-established classification method whose name
refers to the particular training examples, called support vectors, that determine the …

[PDF][PDF] Fast rates in statistical and online learning

T Van Erven, PD Grünwald, NA Mehta, MD Reid… - The Journal of Machine …, 2015 - jmlr.org
The speed with which a learning algorithm converges as it is presented with more data is a
central problem in machine learning—a fast rate of convergence means less data is needed …

[PDF][PDF] Bayesian nonparametric covariance regression

EB Fox, DB Dunson - The Journal of Machine Learning Research, 2015 - jmlr.org
Capturing predictor-dependent correlations amongst the elements of a multivariate
response vector is fundamental to numerous applied domains, including neuroscience …

Support vector machines and linear regression coincide with very high-dimensional features

N Ardeshir, C Sanford, DJ Hsu - Advances in Neural …, 2021 - proceedings.neurips.cc
The support vector machine (SVM) and minimum Euclidean norm least squares regression
are two fundamentally different approaches to fitting linear models, but they have recently …

Randomized learning and generalization of fair and private classifiers: From pac-bayes to stability and differential privacy

L Oneto, M Donini, M Pontil, J Shawe-Taylor - Neurocomputing, 2020 - Elsevier
We address the problem of randomized learning and generalization of fair and private
classifiers. From one side we want to ensure that sensitive information does not unfairly …

PAC-Bayesian analysis of distribution dependent priors: Tighter risk bounds and stability analysis

L Oneto, D Anguita, S Ridella - Pattern Recognition Letters, 2016 - Elsevier
In this paper we bound the risk of the Gibbs and Bayes classifiers (GC and BC), when the
prior is defined in terms of the data generating distribution, and the posterior is defined in …