RICT at the NTCIR-14 QALab- PoliInfo Task

RICT at the NTCIR-14 QALab- PoliInfo Task
复制标题

NTCIR-14 QALab- PoliInfo 任务中的 RICT

DOI:
--
复制
发表时间:
2019
期刊:
--
影响因子:
--
通讯作者:
Kiyohiko Shinomiya
Kiyohiko Shinomiya
中科院分区:
--
文献类型:
--
作者:
Jiawei Yong;Shintaro Kawamura;Katsumi Kanasaki;Shoichi Naitoh;Kiyohiko Shinomiya

文献摘要

参考文献

被引文献

相似文献

我们的RICT团队在NTCIR-14 QA Lab-PoliInfo中处理细分和分类子任务。作为我们在分割任务中的技术特点,我们将片段作为检索对象,利用基于线索阶段的方法,包括诺贝尔半监督学习方法来检测片段边界。在这里,我们提出了5种正式运行的方法。对于分类任务,我们使用不存在主题影响的所有话语来训练我们的有监督学习模型,并利用离群点检测技术来缓解训练数据不均衡的问题。在这里,我们已经提交了7次正式运行。由于评估结果表明我们的方法获得了高于平均水平的分数,所以我们在分类方面所做的主要贡献是提供了一个可行的系统来处理少量训练数据不平衡的问题,特别是在区域大会纪要中。同时,本文的高效方法也为区域大会纪要的分割提供了有效的方法,有助于准确把握区域大会纪要的特征。
Our RICT team tackled segmentation and classification subtasks in NTCIR-14 QA Lab-PoliInfo. As our technical characteristic in segmentation task, we regard segments as retrieval objects, and utilize cue-phase-based methods including a nobel semi-supervised learning method to detect segment boundary. Here we have presented 5 methods for formal run. As to classification task, we train our supervised learning model by all utterances without topic effects and utilize outlier detection technologies to alleviate the imbalance training data problem. Here we have submitted 7 runs for formal run. Since evaluation results show that our approach achieves higher scores than average, the main contribution for classification we made is providing a feasible system to deal with a small number of imbalance training data problem especially in the regional assembly minutes. Meanwhile, we also made contribution for segmentation to grasp distinguishing features of regional assembly minutes by our efficient method.
DOI: 10.1007/978-3-030-36805-0_10
发表时间: 2019-06
期刊: --
影响因子: --
作者:
Yasutomo Kimura;Hideyuki Shibuki;Hokuto Ototake;Yuzu Uchida;Keiichi Takamaru;Kotaro Sakamoto;Madoka Ishioroshi;T. Mitamura;N. Kando;Tatsunori Mori;Harumichi Yuasa;S. Sekine;Kentaro Inui
通讯作者: Yasutomo Kimura;Hideyuki Shibuki;Hokuto Ototake;Yuzu Uchida;Keiichi Takamaru;Kotaro Sakamoto;Madoka Ishioroshi;T. Mitamura;N. Kando;Tatsunori Mori;Harumichi Yuasa;S. Sekine;Kentaro Inui