Online Multiclass Boosting with Bandit Feedback
Online Multiclass Boosting with Bandit Feedback
复制标题
带有 Bandit 反馈的在线多类提升
DOI：
--
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Ambuj Tewari
中科院分区：
文献类型：
--
作者：
Daniel T. Zhang;Young Hun Jung;Ambuj Tewari
我们提出了在线提升算法的多类分类与强盗反馈，学习者只收到反馈的正确性，其预测。我们提出了一个无偏估计的损失使用随机预测，允许模型更新其弱学习有限的信息。使用无偏估计，我们扩展了两个全信息提升算法（Jung等人，2017年，《土匪》上映。我们证明了强盗算法的渐近误差界完全匹配他们的全信息同行。限制反馈的成本反映在更大的样本复杂性中。实验结果也支持我们的理论研究结果，所提出的模型的性能与现有的强盗助推算法，这是有限的使用二进制弱学习器。
We present online boosting algorithms for multiclass classification with bandit feedback, where the learner only receives feedback about the correctness of its prediction. We propose an unbiased estimate of the loss using a randomized prediction, allowing the model to update its weak learners with limited information. Using the unbiased estimate, we extend two full information boosting algorithms (Jung et al., 2017) to the bandit setting. We prove that the asymptotic error bounds of the bandit algorithms exactly match their full information counterparts. The cost of restricted feedback is reflected in the larger sample complexity. Experimental results also support our theoretical findings, and performance of the proposed models is comparable to that of an existing bandit boosting algorithm, which is limited to use binary weak learners.