期刊
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
卷 43, 期 2, 页码 663-678出版社
IEEE COMPUTER SOC
DOI: 10.1109/TPAMI.2019.2932415
关键词
Prediction algorithms; Waste materials; Error analysis; Machine learning; Machine learning algorithms; Inference algorithms; Reliability
资金
- NYU Seed Grant
- NYU WIRELESS
- United States National Science Foundation [CNS-1302336, MCB-1158273, IOS-1339362, MCB-1412232]
SafePredict is a novel meta-algorithm that works with any base prediction algorithm to guarantee a chosen correctness rate by allowing refusals. It does not rely on assumptions about data distribution or base predictor and adapts to changes in the base predictor's error rate without knowing when the changes occur.
SafePredict is a novel meta-algorithm that works with any base prediction algorithm for online data to guarantee an arbitrarily chosen correctness rate, 1 - epsilon, by allowing refusals. Allowing refusals means that the meta-algorithm may refuse to emit a prediction produced by the base algorithm so that the error rate on non-refused predictions does not exceed epsilon. The SafePredict error bound does not rely on any assumptions on the data distribution or the base predictor. When the base predictor happens not to exceed the target error rate epsilon, SafePredict refuses only a finite number of times. When the error rate of the base predictor changes through time SafePredict makes use of a weight-shifting heuristic that adapts to these changes without knowing when the changes occur yet still maintains the correctness guarantee. Empirical results show that (i) SafePredict compares favorably with state-of-the-art confidence-based refusal mechanisms which fail to offer robust error guarantees; and (ii) combining SafePredict with such refusal mechanisms can in many cases further reduce the number of refusals. Our software is included in the supplementary material, which can be found on the Computer Society Digital Library at http://doi.ieeecomputersociety.org/10.1109/TPAMI.2019.2932415.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据