在机器阅读理解任务中,由于中文实词的组合性和隐喻性,且缺乏有关中文实词辨析的数据集,因此传统方法对中文实词的理解程度和辨析能力仍然有限.为此,构建了一个大规模(600k)的中文实词辨析数据集(Chinese notional word discrimination ...在机器阅读理解任务中,由于中文实词的组合性和隐喻性,且缺乏有关中文实词辨析的数据集,因此传统方法对中文实词的理解程度和辨析能力仍然有限.为此,构建了一个大规模(600k)的中文实词辨析数据集(Chinese notional word discrimination cloze data set,CND).在数据集中,一句话中的一个实词被替换成了空白占位符,需要从提供的两个候选实词中选择正确答案.设计了一个基线模型RoBERTa-ND(RoBERTa-based notional word discrimination model)来对候选词进行选择.模型首先利用预训练语言模型提取语境中的语义信息.其次,融合候选实词语义并通过分类任务计算候选词得分.最后,通过增强模型对位置及方向信息的感知,进一步加强了模型的中文实词的辨析能力.实验表明,该模型在CND上准确率达到90.21%,战胜了DUMA(87.59%),GNN-QA(84.23%)等主流的完形填空模型.该工作填补了中文隐喻语义理解研究的空白,可以在提高中文对话机器人认知能力等方向开发更多实用价值.数据集CND及RoBERTa-ND代码均已开源:https://github.com/2572926348/CND-Large-scale-Chinese-National-word-discrimination-dataset.展开更多
This paper treats metaphors and terminology or the discourse of tasawwuf Having in mind that tasawwuf is a discipline developed in the scope of Islamic tradition and that its focus is on the esoteric dimension of the ...This paper treats metaphors and terminology or the discourse of tasawwuf Having in mind that tasawwuf is a discipline developed in the scope of Islamic tradition and that its focus is on the esoteric dimension of the Qur＇an （ta wil）, which relates to the cognition of the highest rank, an approach to such a discourse needs special effort, both intellectual and spiritual. The language of literature belonging to either theoretical or practical tasawwuf is highly symbolic： theoretical tasawwuf deals with ontology and discusses God, world, and human being, while pragmatic tasawwuftries, just as the ethics, to change human being, that is, to ennoble him. Because of its specific way of presentation and its specific discourse, a communication with that discourse is determined by the knowledge of its symbolism, its terminology and, generally, its professional terminology.展开更多
文摘在机器阅读理解任务中,由于中文实词的组合性和隐喻性,且缺乏有关中文实词辨析的数据集,因此传统方法对中文实词的理解程度和辨析能力仍然有限.为此,构建了一个大规模(600k)的中文实词辨析数据集(Chinese notional word discrimination cloze data set,CND).在数据集中,一句话中的一个实词被替换成了空白占位符,需要从提供的两个候选实词中选择正确答案.设计了一个基线模型RoBERTa-ND(RoBERTa-based notional word discrimination model)来对候选词进行选择.模型首先利用预训练语言模型提取语境中的语义信息.其次,融合候选实词语义并通过分类任务计算候选词得分.最后,通过增强模型对位置及方向信息的感知,进一步加强了模型的中文实词的辨析能力.实验表明,该模型在CND上准确率达到90.21%,战胜了DUMA(87.59%),GNN-QA(84.23%)等主流的完形填空模型.该工作填补了中文隐喻语义理解研究的空白,可以在提高中文对话机器人认知能力等方向开发更多实用价值.数据集CND及RoBERTa-ND代码均已开源:https://github.com/2572926348/CND-Large-scale-Chinese-National-word-discrimination-dataset.
文摘This paper treats metaphors and terminology or the discourse of tasawwuf Having in mind that tasawwuf is a discipline developed in the scope of Islamic tradition and that its focus is on the esoteric dimension of the Qur＇an （ta wil）, which relates to the cognition of the highest rank, an approach to such a discourse needs special effort, both intellectual and spiritual. The language of literature belonging to either theoretical or practical tasawwuf is highly symbolic： theoretical tasawwuf deals with ontology and discusses God, world, and human being, while pragmatic tasawwuftries, just as the ethics, to change human being, that is, to ennoble him. Because of its specific way of presentation and its specific discourse, a communication with that discourse is determined by the knowledge of its symbolism, its terminology and, generally, its professional terminology.
