2018

Extreme Value Theory for Open Set Classification -- GPD and GEV Classifiers

Vignotto, Edoardo, Engelke, Sebastian

Understand

Classification tasks usually assume that all possible classes are present during the training phase.

This is restrictive if the algorithm is used over a long time and possibly encounters samples from unknown classes.
The recently introduced extreme value machine, a classifier motivated by extreme value theory, addresses this problem and achieves competitive performance in specific cases.
We show that this algorithm can fail when the geometries of known and unknown classes differ.

Built on

On estimating the endpoint of a distribution
P. Hall · 1982
Earlier work this paper cites.
Inductive knowledge acquisition: a case study
J. R. Quinlan, P. J. Compton, K. A. Horn, and L. Lazarus · 1986
Earlier work this paper cites.
Letter recognition using holland-style adaptive classifiers
P. W. Frey and D. J. Slate · 1991
Earlier work this paper cites.
Synthesis and performance analysis of multilayer neural network architectures
W. Schiffmann, M. Joost, and R. Werner · 1992
Earlier work this paper cites.
Novelty detection and neural network validation
C. M. Bishop · 1994
Earlier work this paper cites.
Multiple hypothesis testing
J. P. Shaffer · 1995
Earlier work this paper cites.
The use of the area under the ROC curve in the evaluation of machine learning algorithms
A. P. Bradley · 1997
Earlier work this paper cites.
An optimal algorithm for approximate nearest neighbor searching fixed dimensions
S. Arya, D. M. Mount, N. S. Netanyahu, R. Silverman, and A. Y. Wu · 1998
Earlier work this paper cites.
Novelty detection using extreme value statistics
S. J. Roberts · 1999
Earlier work this paper cites.
Support vector method for novelty detection
B. Schölkopf, R. C. Williamson, A. J. Smola, J. Shawe-Taylor, and J. C. Platt · 2000
Earlier work this paper cites.
An introduction to statistical modeling of extreme values
S. Coles, J. Bawa, L. Trenner, and P. Dorazio · 2001
Earlier work this paper cites.
Incremental learning with support vector machines
S. Ruping · 2001
Earlier work this paper cites.
Using AUC and accuracy in evaluating learning algorithms
J. Huang and C. X. Ling · 2005
Earlier work this paper cites.

Similar

Outlier detection by active learning
N. Abe, B. Zadrozny, and J. Langford · 2006
Cited alongside, same era.
A review of statistical outlier methods
S. Walfish · 2006
Cited alongside, same era.
Extreme value theory: an introduction
L. De Haan and A. Ferreira · 2007
Cited alongside, same era.
A hybrid machine learning approach to network anomaly detection
T. Shon and J. Moon · 2007
Cited alongside, same era.
ImageNet: A Large-Scale Hierarchical Image Database
J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei · 2009
Cited alongside, same era.
Learning multiple layers of features from tiny images
Alex Krizhevsky and Geoffrey Hinton · 2009
Cited alongside, same era.
On-line random forests
A. Saffari, C. Leistner, J. Santner, M. Godec, and H. Bischof · 2009
Cited alongside, same era.
Meta-recognition: the theory and practice of recognition score analysis
W. J. Scheirer, A. Rocha, R. J. Micheals, and T. E. Boult · 2011
Cited alongside, same era.
Metric learning for large scale image classification: Generalizing to new classes at near-zero cost
T. Mensink, J. Verbeek, F. Perronnin, and G. Csurka · 2012
Cited alongside, same era.
One class random forests
C. Désir, S. Bernard, C. Petitjean, and L. Heutte · 2013
Cited alongside, same era.
EVSAC: accelerating hypotheses generation by modeling matching scores with extreme value theory
V. Fragoso, P. Sen, S. Rodriguez, and M. Turk · 2013
Cited alongside, same era.
Speech recognition with deep recurrent neural networks
A. Graves, A. Mohamed, and G. Hinton · 2013
Cited alongside, same era.

Then

A review of novelty detection
M. A. F. Pimentel, D. A. Clifton, L. Clifton, and L. Tarassenko · 2014
Later among the works it cites.
Probability models for open set recognition
W. J. Scheirer, L. P. Jain, and T. E. Boult · 2014
Later among the works it cites.
Towards open world recognition
A. Bendale and T. Boult · 2015
Later among the works it cites.
Deep neural networks are easily fooled: high confidence predictions for unrecognizable images
A. Nguyen, J. Yosinski, and J. Clune · 2015
Later among the works it cites.
Pattern Recognition and Machine Learning
M. B. Christopher · 2016
Later among the works it cites.
Deep residual learning for image recognition
K. He, X. Zhang, S. Ren, and J. Sun · 2016
Later among the works it cites.
UCI machine learning repository, 2017
D. Dua and C. Graff · 2017
Later among the works it cites.
Estimation of extreme depth-based quantile regions
Yi He and John H. J. Einmahl · 2017
Later among the works it cites.
iCaRL: Incremental classifier and representation learning
S. Rebuffi, A. Kolesnikov, and C. H. Lampert · 2017
Later among the works it cites.
Extreme value theory-based methods for visual recognition
W. J. Scheirer · 2017
Later among the works it cites.
Recent advances in open set recognition: A survey
Original
C. Geng, S. Huang, and S. Chen · 2018
Closest in time.
On binary classification in extreme regions
H. Jalalzai, S. Clémençon, and A. Sabourin · 2018
Closest in time.
The extreme value machine
E. M. Rudd, L. P. Jain, W. J. Scheirer, and T. E. Boult · 2018
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…