A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Filters
Benign Overfitting in Multiclass Classification: All Roads Lead to Interpolation
[article]
2022
arXiv
pre-print
Preserved Fulltext
Other Versions
Motivated by this discrepancy, we study benign overfitting in multiclass linear classification. ...
The literature on "benign overfitting" in overparameterized models has been mostly restricted to regression or binary classification; however, modern machine learning operates in the multiclass setting ...
extent the phenomena of benign overfitting and double descent [BHMM19, GJS + 20] can be proven to occur in classification tasks. ...
arXiv:2106.10865v2
fatcat:d42gdf5dtvb3vh7rj4ytvgsn54
Network Attack Detection Method of the Cyber-Physical Power System Based on Ensemble Learning
2022
Applied Sciences
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Network attack, in addition to faults, becomes an important factor restricting the stable operation of the power system. ...
At the same time, in the process of gradient boost, the focal loss was introduced to optimize the attention weight of the classifier to the misclassified samples, thus improving the network attack detection ...
may lead to high rates of false alarms. ...
doi:10.3390/app12136498
fatcat:jqv2tbea5rerfb2jcfbts5gdbm
Identifying and Exploiting Structures for Reliable Deep Learning
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
To do this, we identify structures in deep neural networks that can be exploited to mitigate the above causes of unreliability of deep learning algorithms. ...
However, as recent works point out, these systems suffer from several issues that make them unreliable for use in the real world, including vulnerability to adversarial attacks (Szegedy et al. [248]), ...
blow up, a phenomenon referred to as benign overfitting. ...
arXiv:2108.07083v1
fatcat:lducrn5tlfeqvpxevz6gukfvse
From Dependence to Causation
[article]
2016
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Second, we build on this framework to interpret the problem of causal inference as the task of distribution classification, yielding a family of novel causal inference algorithms. ...
But, currently available causal inference algorithms operate in specific regimes, and rely on assumptions that are difficult to verify in practice. ...
Naïvely, one could extend the framework presented in Section 7.2 from the binary classification of 2-dimensional distributions to the multiclass classification of d-dimensional distributions. ...
arXiv:1607.03300v1
fatcat:img5m23n5ncx5mfejgqkjft2ua
Semantic Segmentation of Ambiguous Images
2020
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Rauschen in der Aufnahme. ...
Ähnliche Szenarien existieren in natürlichen Bildern, in welchen die Kontextinformation, die es braucht um Mehrdeutigkeiten aufzulösen, limitiert sein kann, beispielsweise aufgrund von Verdeckungen oder ...
All patients had a clinical indication leading to prostate biopsy (details on the biopsy protocol are given in Sec. ...
doi:10.5445/ir/1000118113
fatcat:ey6bbnzdgngl7lpumxgyn337le
New Paradigms and Optimality Guarantees in Statistical Learning and Estimation
2018
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
overfitting on public benchmarks. ...
Moreover, the "common task framework" that is adopted by many sub- disciplines of AI has made it possible for many people to collaboratively and repeated work on the same data set, leading to implicit ...
For the test itself, we consider our kth order KS test, in ...
doi:10.1184/r1/6720836.v1
fatcat:bdtsjzawkbdjxntc5pmm5irvna
Dagstuhl Reports, Volume 6, Issue 11, November 2016, Complete Issue
[article]
2017
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
potential functions methods, in SVM, the general "kernel trick", the observation that kernels can be defined on arbitrary sets of objects, the link to GPs, and finally the idea to represent distributions ...
Kernel mean representations lend themselves well to the development of kernel methods for probabilistic programming, i.e., methods for lifting functional operations defined for data types to the same functional ...
, classification of data according to their impact on real world. ...
doi:10.4230/dagrep.6.11
fatcat:tfkdfittpjdydfv7ejvk4bvnh4