A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2015; you can also visit the original URL.
The file type is application/pdf
.
Filters
A minimax classification approach with application to robust speech recognition
1993
IEEE Transactions on Speech and Audio Processing
Abstruct-A minimax approach for robust classification of parametric information sources is studied and applied to isolatedword speech recognition based on hidden Markov modeling. ...
The goal is to reduce the sensitivity of speech recognition systems to a possible mismatch between the training and testing conditions. ...
ACKNOWLEDGMENT The authors are grateful to Yariv Ephraim for providing us recognition scores from his gain adapted speech recognition system. ...
doi:10.1109/89.221371
fatcat:e3lu2lndvrfd7hyovi2pkfojmu
A Minimax Classification Approach With Application To Robust Speech Recognition
Proceedings. 1991 IEEE International Symposium on Information Theory
Abstruct-A minimax approach for robust classification of parametric information sources is studied and applied to isolatedword speech recognition based on hidden Markov modeling. ...
The goal is to reduce the sensitivity of speech recognition systems to a possible mismatch between the training and testing conditions. ...
ACKNOWLEDGMENT The authors are grateful to Yariv Ephraim for providing us recognition scores from his gain adapted speech recognition system. ...
doi:10.1109/isit.1991.695341
fatcat:erzhkj6renhpvh2ezp3ffumt7y
Adversarially Robust Classification based on GLRT
[article]
2020
arXiv
pre-print
We show that the GLRT approach yields performance competitive with that of the minimax approach under the worst-case attack, and observe that it yields a better robustness-accuracy trade-off under weaker ...
We evaluate the GLRT approach for the special case of binary hypothesis testing in white Gaussian noise under ℓ_∞ norm-bounded adversarial perturbations, a setting for which a minimax strategy optimizing ...
CONCLUSION The GLRT approach to robust hypothesis testing explored in this paper can be generalized to complex models, in contrast to the difficulty of finding optimal minimax classifiers. ...
arXiv:2011.07835v1
fatcat:wysgs7c745gx7jiyljmkvxrf7u
Robust classification by a nearest mean-median rule for generalized Gaussian pattern distributions
2008
Pattern Recognition and Image Analysis
To provide stability of classification, a robust supervised minimum distance classifier based on the minimax (in the Huber sense) estimate of location is designed for the class of generalized Gaussian ...
pattern distributions with a bounded variance. ...
Finally, we note that the proposed robust classification rules can be extended to the multivariate case on the basis of a coordinate-wise approach. ------------------------------------------------, --- ...
doi:10.1134/s1054661808020107
fatcat:hepjcdpuzbhhlbirwbxvub63cy
Robust Learning in Heterogeneous Contexts
[article]
2022
arXiv
pre-print
We develop a robust method that takes into account the uncertainty of the context distribution. ...
Unlike the conventional and overly conservative minimax approach, we focus on excess risks and construct distribution sets with statistical coverage to achieve an appropriate trade-off between performance ...
For instance, in a classification task the minimax approach could focus on a context where the error rate is high but insensitive with respect to θ, while ignoring contexts with low minimum error rates ...
arXiv:2105.08532v3
fatcat:fzdyzffrbfabdidqb3cxy5kc3u
Minimax Regret Classifier for Imprecise Class Distributions
2007
Journal of machine learning research
A neural-based minimax regret classifier for general multi-class decision problems is presented. Experimental results show its robustness and the advantages in relation to other approaches. ...
In this paper we propose a minimax regret (minimax deviation) approach, that seeks to minimize the maximum deviation from the performance of the optimal risk classifier. ...
We seek for a system as robust as the conventional minimax approach but less pessimistic at the same time. We will refer to it as a minimax deviation (or minimax regret) classifier. ...
dblp:journals/jmlr/Alaiz-RodriguezGC07
fatcat:e6hjpani75h5llbht7zv3tlply
Generalized Likelihood Ratio Test for Adversarially Robust Hypothesis Testing
[article]
2021
arXiv
pre-print
For non-asymptotic regimes, we show via simulations that the GLRT defense is competitive with the minimax approach under the worst-case attack, while yielding a better robustness-accuracy tradeoff under ...
We also illustrate the GLRT approach for a multi-class hypothesis testing problem, for which a minimax strategy is not known, evaluating its performance under both noise-agnostic and noise-aware adversarial ...
Optimal robust
approach to multi-class settings for which minimax strate- classifiers are derived for binary and ternary classification
gies are not known. ...
arXiv:2112.02209v1
fatcat:7lh3gixnd5e7zh3b44qnwiyxdy
Page 6449 of Mathematical Reviews Vol. , Issue 2003h
[page]
2003
Mathematical Reviews
ISBN 1-4020-0805-8
Publisher's description: “This book discusses a new approach to the classification problem following the decision support orienta- tion of multicriteria decision aid. ...
The robust character of minimax, mentioned above, is central to the usefulness of the strategies discussed in this book. ...
Affine Distortion Compensation for an Isolated Online Handwritten Chinese Character Using Combined Orientation Estimation and HMM-Based Minimax Classification
2009
2009 10th International Conference on Document Analysis and Recognition
Depending on the number of possible orientations, an HMM-based minimax classification approach is then used to estimate an affine transformation against either the original sample or the compensated sample ...
This paper presents a new approach to compensating affine distortion of an isolated online handwritten Chinese character. ...
To speed up the estimation process, the minimax classification is conducted on a short-list of candidates hypothesized by an efficient single-prototype (SP) based classifier which is robust against rotation ...
doi:10.1109/icdar.2009.87
dblp:conf/icdar/HeH09a
fatcat:tjgbyomp7bdjberfck4c7lye3a
Robust Classification Under Sample Selection Bias
2014
Neural Information Processing Systems
We develop a framework for learning a robust bias-aware (RBA) probabilistic classifier that adapts to different sample selection biases using a minimax estimation formulation. ...
We demonstrate the behavior and effectiveness of our approach on binary classification tasks. ...
We propose a novel approach to classification that embraces the uncertainty resulting from sample selection bias by producing predictions that are explicitly robust to it. ...
dblp:conf/nips/LiuZ14
fatcat:z2p546qgfjfh7ojy6zinxofyry
A Maxmin Approach to Optimize Spatial Filters for EEG Single-Trial Classification
[chapter]
2009
Lecture Notes in Computer Science
This work contributes by developing a minimax approach to robustify the common spatial patterns (CSP) algorithm. ...
EEG single-trial analysis requires methods that are robust with respect to noise, artifacts and nonstationarity among other problems. ...
[8] applied a minimax approach to Fisher Discrimant Analysis (FDA) and proposed a novel robust classification method. ...
doi:10.1007/978-3-642-02478-8_84
fatcat:lfxb64zllnfitkvgd4u4onzdna
A minimax search algorithm for robust continuous speech recognition
2000
IEEE Transactions on Speech and Audio Processing
ACKNOWLEDGMENT The authors would like to thank Dr. C.-H. Lee at Bell Laboratories for his comments and discussions on this work. ...
This scheme becomes a potential approach for robust ASR because no rigid assumptions about the sources and mechanisms of the mismatches have to be made. ...
Because of its intrinsic nature of a recursive search, the approach can be easily extended to perform CSR. ...
doi:10.1109/89.876302
fatcat:c2mfxt2zxzda7kiarav7kdegty
Minimax Robust MIMO Radar Waveform Design
2007
IEEE Journal on Selected Topics in Signal Processing
Our findings also indicate that the MI and MMSE criteria lead to different minimax robust waveform solutions, which is in contrast to the case of the perfectly known target PSD. ...
and classification. ...
Using a minimax robust design optimizes the worst-case performance, and is similar to designing for the worst case which is a well accepted engineering approach. ...
doi:10.1109/jstsp.2007.897056
fatcat:pjulzr7u45c4jktowax5j2wwqu
A Minimax Approach to Supervised Learning
[article]
2017
arXiv
pre-print
Applying this principle to sets of distributions with marginal on X constrained to be the empirical marginal from the data, we develop a general minimax approach for supervised learning problems. ...
We also prove a bound on the generalization worst-case error in the minimax approach. ...
Acknowledgments We are grateful to Stanford University providing a Stanford Graduate Fellowship, and the Center for Science of Information (CSoI), an NSF Science and Technology Center under grant agreement ...
arXiv:1606.02206v5
fatcat:mhtn2yma5bcizk2gu26cc4zs6u
MRCpy: A Library for Minimax Risk Classifiers
[article]
2024
arXiv
pre-print
We present a Python library, MRCpy, that implements minimax risk classifiers (MRCs) based on the robust risk minimization (RRM) approach. ...
MRCpy follows an object-oriented approach and adheres to the standards of popular Python libraries, such as scikit-learn, facilitating readability and easy usage together with a seamless integration with ...
A basic approach to implement a new feature mapping can be to extend BasePhi and implement the functions fit and transform. ...
arXiv:2108.01952v4
fatcat:qrwrs2opl5ccteebamraiqnmmm
« Previous
Showing results 1 — 15 out of 6,812 results