Aligning model representations to humans has been found to improve robustness and generalization. However, such methods often focus on standard observational data. Synthetic data is proliferating and powering many advances in machine learning; yet, it is not always clear whether synthetic labels are perceptually aligned to humans -- rendering it likely model representations are not human aligned. We focus on the synthetic data used in mixup: a powerful regularizer shown to improve model robustness, generalization, and calibration. We design a comprehensive series of elicitation interfaces, which we release as HILL MixE Suite, and recruit 159 participants to provide perceptual judgments along with their uncertainties, over mixup examples. We find that human perceptions do not consistently align with the labels traditionally used for synthetic points, and begin to demonstrate the applicability of these findings to potentially increase the reliability of downstream models, particularly when incorporating human uncertainty. We release all elicited judgments in a new data hub we call H-Mix.
翻译:对人类的模型表达方式进行了调整,以提高稳健性和普遍性。然而,这些方法往往侧重于标准观测数据。合成数据正在扩散,并给机器学习方面的许多进步提供动力;然而,尚不总是清楚合成标签是否与人类感知一致 -- -- 使模型表达方式可能与人类不相协调。我们侧重于混合中使用的合成数据:显示一个强大的常规化器,以提高模型的稳健性、普遍性和校准性。我们设计了一系列全面的吸引界面,我们以HILL MixE 套件的形式发布,并征聘159名参与者,提供概念判断及其不确定性的判断,并超越混合范例。我们发现,人类的认知与传统上用于合成点的标签不一致,并开始表明这些结论的可适用性,以可能提高下游模型的可靠性,特别是在纳入人类不确定性时。我们都在一个新的数据枢纽中发布所有引出的判决,我们称之为H-Mix。</s>