arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.03425eess.AScs.CLcs.LGcs.SD
收藏

端到端语音理解中的深度F测度最大化

Deep F-measure Maximization for End-to-End Speech Understanding

University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)

Leda Sarı, Mark Hasegawa-Johnson

摘要:

口语理解(SLU)数据集与许多其他机器学习数据集一样,通常会遇到标签不平衡问题。标签不平衡通常会导致学习模型在输出中复制类似的偏差,这会引起对数据集中少数类的不公平问题。在这项工作中,我们通过最大化F度量而不是神经网络模型训练中的准确性来解决公平性问题。我们提出了一个可微近似的F-措施和训练网络与此目标使用标准的反向传播。我们在两个标准的公平数据集上进行实验,成人,社区和犯罪,以及ATIS数据集上的语音到意图检测和Speech-COCO数据集上的语音到图像概念分类。在所有这四个任务中,F-measure最大化导致微F1分数的改善,与使用交叉熵损失函数训练的模型相比,绝对改善高达8%。在两个多类SLU任务中,所提出的方法显著提高了类覆盖率,即,积极回忆的班级数量。

英文摘要:

Spoken language understanding (SLU) datasets, like many other machine learning datasets, usually suffer from the label imbalance problem. Label imbalance usually causes the learned model to replicate similar biases at the output which raises the issue of unfairness to the minority classes in the dataset. In this work, we approach the fairness problem by maximizing the F-measure instead of accuracy in neural network model training. We propose a differentiable approximation to the F-measure and train the network with this objective using standard backpropagation. We perform experiments on two standard fairness datasets, Adult, and Communities and Crime, and also on speech-to-intent detection on the ATIS dataset and speech-to-image concept classification on the Speech-COCO dataset. In all four of these tasks, F-measure maximization results in improved micro-F1 scores, with absolute improvements of up to 8% absolute, as compared to models trained with the cross-entropy loss function. In the two multi-class SLU tasks, the proposed approach significantly improves class coverage, i.e., the number of classes with positive recall.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑