Theoretical Computer Science ( IF 0.9 ) Pub Date : 2020-08-27 , DOI: 10.1016/j.tcs.2020.08.025 Arnab Bhattacharyya , Ameet Gadekar , Ninad Rajgopal
We consider the problem of learning k-parities in the online mistake-bound model: given a hidden vector where the hamming weight of x is k and a sequence of “questions” , where the algorithm must reply to each question with , what is the best trade-off between the number of mistakes made by the algorithm and its time complexity? We improve the previous best result of Buhrman et al. [3] by an factor in the time complexity.
Next, we consider the problem of learning k-parities in the PAC model in the presence of random classification noise of rate . Here, we observe that even in the presence of classification noise of non-trivial rate, it is possible to learn k-parities in time better than , whereas the current best algorithm for learning noisy k-parities, due to Grigorescu et al. [9], inherently requires time even when the noise rate is polynomially small.
中文翻译:
改进的k奇偶性学习
我们考虑在线错误约束模型中学习k-奇偶性的问题:给定一个隐藏向量x的汉明权重为k以及一系列“问题”,算法必须在其中回答每个问题 ,算法所犯的错误数量与其时间复杂度之间的最佳权衡是什么?我们提高了Buhrman等人先前的最佳结果。[3]由 时间复杂度的因素。
接下来,我们考虑在速率随机分类噪声存在下在PAC模型中学习k-奇偶性的问题。在这里,我们看到,即使是在不平凡率的分类噪声的存在,有可能学习ķ -parities在时间上比好,而由于Grigorescu等人,目前用于学习有噪k奇偶校验的最佳算法。[9],本质上需要时间 即使噪声率在多项式上很小。