A Permutation-Based Model for Crowd Labeling: Optimal Estimation and Robustness

A Permutation-Based Model for Crowd Labeling: Optimal Estimation and Robustness
复制标题

基于排列的人群标记模型:最优估计和鲁棒性

DOI:
10.1109/tit.2020.3045613
复制
发表时间:
2016
影响因子:
2.5
通讯作者:
M. Wainwright
M. Wainwright
中科院分区:
计算机科学2区
文献类型:
--
作者:
Nihar B. Shah;Sivaraman Balakrishnan;M. Wainwright

文献摘要

参考文献

被引文献

相似文献

随着众包平台和大规模数据集的出现,聚合和去噪众包标记数据的任务变得越来越重要。我们提出了一个基于置换的人群标记数据模型,这是一个显着的推广经典Dawid-Skene模型,并引入了一个新的误差度量,通过它来比较不同的估计。我们推导出全局极小极大速率的置换为基础的模型,是尖锐的对数因子,并匹配的极小极大下界下得到的更简单的Dawid-Skene模型。然后,我们设计了两个计算效率高的估计:WAN估计的设置,其中工人的能力方面的排序是近似已知的,和OBI-WAN估计,这是未知的。对于每一个这些估计,我们提供了非渐近界的性能。我们对真实世界的众包数据进行了合成模拟和实验,实验结果证实了我们的理论研究结果。
The task of aggregating and denoising crowd-labeled data has gained increased significance with the advent of crowdsourcing platforms and massive datasets. We propose a permutation-based model for crowd labeled data that is a significant generalization of the classical Dawid-Skene model, and introduce a new error metric by which to compare different estimators. We derive global minimax rates for the permutation-based model that are sharp up to logarithmic factors, and match the minimax lower bounds derived under the simpler Dawid-Skene model. We then design two computationally-efficient estimators: the WAN estimator for the setting where the ordering of workers in terms of their abilities is approximately known, and the OBI- WAN estimator where that is not known. For each of these estimators, we provide non-asymptotic bounds on their performance. We conduct synthetic simulations and experiments on real-world crowdsourcing data, and the experimental results corroborate our theoretical findings.
DOI: --
发表时间: 2019-12
期刊: --
影响因子: --
作者:
Ivan Stelmakh;Nihar B. Shah;Aarti Singh
通讯作者: Ivan Stelmakh;Nihar B. Shah;Aarti Singh
DOI: 10.3150/17-bej1000
发表时间: 2019-02-01
期刊: BERNOULLI
影响因子: 1.5
作者:
Flammarion, Nicolas;Mao, Cheng;Rigollet, Philippe
通讯作者: Rigollet, Philippe
DOI: --
发表时间: 2018-06
期刊: J. Mach. Learn. Res.
影响因子: --
作者:
Ivan Stelmakh;Nihar B. Shah;Aarti Singh
通讯作者: Ivan Stelmakh;Nihar B. Shah;Aarti Singh
你的 2 是我的 1,你的 3 是我的 9:处理评级中的任意错误校准
DOI: --
发表时间: 2019
期刊: AAMAS Conference proceedings
影响因子: --
作者:
Wang, J;Shah, N
通讯作者: Shah, N