Search Machine Learning Repository: @inproceedings{icml2013_xu13b,
    Publisher = {JMLR Workshop and Conference Proceedings},
    Title = {Anytime Representation Learning},
    Url = {},
    Booktitle = {Proceedings of the 30th International Conference on Machine Learning (ICML-13)},
    Author = {Zhixiang Xu and Matt Kusner and Gao Huang and Kilian Q. Weinberger},
    Number = {3},
    Month = may,
    Volume = {28},
    Editor = {Sanjoy Dasgupta and David Mcallester},
    Year = {2013},
    Pages = {1076-1084},
    Abstract = {Evaluation cost during test-time is becoming increasingly important as many real-world applications need fast evaluation (e.g. web search engines, email spam filtering) or use expensive features (e.g. medical diagnosis). We introduce Anytime Feature Representations (AFR), a novel algorithm that explicitly addresses this trade-off in the data representation rather than in the classifier. This enables us to turn conventional classifiers, in particular Support Vector Machines, into test-time cost sensitive anytime classifiers combining the advantages of anytime learning and large-margin classification.}