• Corpus ID: 204576058

Elementos da teoria de aprendizagem de máquina supervisionada

  title={Elementos da teoria de aprendizagem de m{\'a}quina supervisionada},
  author={Vladimir G. Pestov},
  • V. Pestov
  • Published 6 October 2019
  • Computer Science, Mathematics
  • ArXiv
This is a set of lecture notes for an introductory course (advanced undergaduates or the 1st graduate course) on foundations of supervised machine learning (in Portuguese). The topics include: the geometry of the Hamming cube, concentration of measure, shattering and VC dimension, Glivenko-Cantelli classes, PAC learnability, universal consistency and the k-NN classifier in metric spaces, dimensionality reduction, universal approximation, sample compression. There are appendices on metric and… 

On the Vapnik-Chervonenkis dimension of products of intervals in $\mathbb{R}^d$

It is concluded that the VapnikChervonenkis dimension of the set of balls in l ∞ – which denotes R equipped with the sup norm – equals ⌊(3d+ 1)/2⌋.



Scale-sensitive dimensions, uniform convergence, and learnability

A characterization of learnability in the probabilistic concept model, solving an open problem posed by Kearns and Schapire, and shows that the accuracy parameter plays a crucial role in determining the effective complexity of the learner's hypothesis class.

Distance Metric Learning for Large Margin Nearest Neighbor Classification

This paper shows how to learn a Mahalanobis distance metric for kNN classification from labeled examples in a globally integrated manner and finds that metrics trained in this way lead to significant improvements in kNN Classification.

Borel Isomorphic Dimensionality Reduction of Data and Supervised Learning

In this project we further investigate the idea of reducing the dimensionality of datasets using a Borel isomorphism with the purpose of subsequently applying supervised learning algorithms, as

Intrinsic dimension of a dataset: what properties does one expect?

  • V. Pestov
  • Computer Science
    2007 International Joint Conference on Neural Networks
  • 2007
An axiomatic approach to the concept of an intrinsic dimension of a dataset, based on a viewpoint of geometry of high-dimensional structures, postulates that high values of dimension be indicative of the presence of the curse of dimensionality.

Sample Compression, Learnability, and the Vapnik-Chervonenkis Dimension

It is demonstrated that the existence of a sample compression scheme of fixed-size for aclass C is sufficient to ensure that the class C is pac-learnable, and the relationship between sample compression schemes and the VC dimension is explored.

A Compression Approach to Support Vector Model Selection

Inspired by several generalization bounds, "compression coefficients" for SVMs are constructed which measure the amount by which the training labels can be compressed by a code built from the separating hyperplane and can fairly accurately predict the parameters for which the test error is minimized.

PAC learnability under non-atomic measures: A problem by Vidyasagar

An axiomatic approach to intrinsic dimension of a dataset

Sample compression schemes for VC classes

It is shown that every concept class C with VC dimension d has a sample compression scheme of size exponential in d, and an approximate minimax phenomenon for binary matrices of low VC dimension is used, which may be of interest in the context of game theory.