Deep-learning real/bogus classification for the Tomo-e Gozen transient survey
Deep-learning real/bogus classification for the Tomo-e Gozen transient survey
复制标题
DOI：
10.1093/pasj/psac047
复制
发表时间：
2022-06
影响因子：
2.3
通讯作者：
I. Takahashi;R. Hamasaki;N. Ueda;Masaomi Tanaka;N. Tominaga;S. Sako;R. Ohsawa;N. Yoshida
中科院分区：
文献类型：
--
作者：
I. Takahashi;R. Hamasaki;N. Ueda;Masaomi Tanaka;N. Tominaga;S. Sako;R. Ohsawa;N. Yoshida
我们提出了一个深度神经网络真实的/伪分类器，通过处理训练数据中的标签错误来提高Tomo-e Gozen Transient调查中的分类性能。在Tomo-e Gozen的宽场高频瞬态调查中，常规卷积神经网络分类器的性能不足，因为每晚出现约106个虚假检测。为了得到更好的分类器，我们提出了一种新的两阶段训练方法。在该训练方法中，训练数据中的标签错误首先通过正常的监督学习分类检测，然后将它们去除标记并用于半监督学习的训练。对于实际观察到的数据，采用该方法的分类器在0.9的真阳性率（TPR）下实现了0.9998的曲线下面积（AUC）和0.0002的假阳性率（FPR）。这种训练方法节省了人工重新标记的工作，并且在具有高比例标签错误的训练数据上效果更好。通过在Tomo-e Gozen流水线中实现开发的分类器，瞬态候选对象的数量减少到每晚2040个对象，这是以前版本的201/130，同时保持真实的瞬态的恢复率。这使得能够更有效地选择用于后续观察的目标。
We present a deep neural network real/bogus classifier that improves classification performance in the Tomo-e Gozen Transient survey by handling label errors in the training data. In the wide-field, high-frequency transient survey with Tomo-e Gozen, the performance of conventional convolutional neural network classifiers is not sufficient as about 106 bogus detections appear every night. In need of a better classifier, we have developed a new two-stage training method. In this training method, label errors in the training data are first detected by normal supervised learning classification, and then they are unlabeled and used for training of semi-supervised learning. For actual observed data, the classifier with this method achieves an area under the curve (AUC) of 0.9998 and a false positive rate (FPR) of 0.0002 at a true positive rate (TPR) of 0.9. This training method saves relabeling effort by humans and works better on training data with a high fraction of label errors. By implementing the developed classifier in the Tomo-e Gozen pipeline, the number of transient candidates was reduced to ∼40 objects per night, which is ∼1/130 of the previous version, while maintaining the recovery rate of real transients. This enables more efficient selection of targets for follow-up observations.