Open Problem: How fast can a multiclass test set be overfit?

Vitaly Feldman, Roy Frostig, Moritz Hardt
Proceedings of the Thirty-Second Conference on Learning Theory, PMLR 99:3185-3189, 2019.

Abstract

We ask how many measurements of the accuracy on a multiclass benchmark are needed to achieve a given amount of overfitting.

Cite this Paper

BibTeX
@InProceedings{pmlr-v99-feldman19b, title = {Open Problem: How fast can a multiclass test set be overfit?}, author = {Feldman, Vitaly and Frostig, Roy and Hardt, Moritz}, booktitle = {Proceedings of the Thirty-Second Conference on Learning Theory}, pages = {3185--3189}, year = {2019}, editor = {Beygelzimer, Alina and Hsu, Daniel}, volume = {99}, series = {Proceedings of Machine Learning Research}, month = {25--28 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v99/feldman19b/feldman19b.pdf}, url = {https://proceedings.mlr.press/v99/feldman19b.html}, abstract = {We ask how many measurements of the accuracy on a multiclass benchmark are needed to achieve a given amount of overfitting.} }
Endnote
%0 Conference Paper %T Open Problem: How fast can a multiclass test set be overfit? %A Vitaly Feldman %A Roy Frostig %A Moritz Hardt %B Proceedings of the Thirty-Second Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2019 %E Alina Beygelzimer %E Daniel Hsu %F pmlr-v99-feldman19b %I PMLR %P 3185--3189 %U https://proceedings.mlr.press/v99/feldman19b.html %V 99 %X We ask how many measurements of the accuracy on a multiclass benchmark are needed to achieve a given amount of overfitting.
APA
Feldman, V., Frostig, R. & Hardt, M.. (2019). Open Problem: How fast can a multiclass test set be overfit?. Proceedings of the Thirty-Second Conference on Learning Theory, in Proceedings of Machine Learning Research 99:3185-3189 Available from https://proceedings.mlr.press/v99/feldman19b.html.

Related Material