A Study of Distributed Representations for Figures of Research Articles

A Study of Distributed Representations for Figures of Research Articles
复制标题

研究文章图形的分布式表示研究

DOI:
10.1007/978-3-030-72113-8_19
复制
发表时间:
2021
期刊:
ECIR 2021: Advances in Information Retrieval pp. 284-297
影响因子:
--
通讯作者:
Kuzi, Saar
Kuzi, Saar
中科院分区:
--
文献类型:
--
作者:
Kuzi, Saar

文献摘要

参考文献

相似文献

研究文章中的图形是可以直接用于许多应用系统以帮助研究人员的实体,这使得图形的表示成为一个值得研究的问题。在本文中,我们研究了使用深度神经网络学习的数字分布式表示的有效性。我们使用文本和图像数据来学习表示,并比较不同的模型架构和任务的损失函数。此外,为了克服任务缺乏训练数据的问题,我们提出并研究了一种新的弱监督方法来学习嵌入向量,并证明了它比使用最近的工作所建议的一些预训练的神经模型更有效。使用ACL选集中的数字的实验结果表明,研究数字的分布式表示可以比以前研究的词袋表示更有效。然而,结合这两种方法可以进一步提高性能。最后,结果还表明,这些表示,虽然有效的一般,可以敏感的学习方法,使用图像数据和文本和一个简单的模型架构是最有效的方法。
Figures of research articles are entities that can be directly used in many application systems to assist researchers, making the representation of figures a problem worth studying. In this paper, we study the effectiveness of distributed representations, learned using deep neural networks, for figures. We learn representations using both text and image data and compare different model architectures and loss functions for the task. Furthermore, to overcome the lack of training data for the task, we propose and study a novel weak supervision approach for learning embedding vectors and show that it is more effective than using some of the pre-trained neural models as suggested by recent works. Experimental results using figures from the ACL Anthology show that distributed representations for research figures can be more effective than the previously studied bag-of-words representations. Yet, combining the two approaches can further improve performance. Finally, the results also show that these representations, while effective in general, can be sensitive to the learning approach used and that using both image data and text and a simple model architecture is the most effective approach.
DOI: 10.1016/j.jbi.2011.05.003
发表时间: 2011-10
影响因子: 4.5
作者:
Kim, Daehyun;Ramesh, Balaji Polepalli;Yu, Hong
通讯作者: Yu, Hong
DOI: --
发表时间: 2019
期刊: European Conference on Information Retrieval
影响因子: --
作者:
Saar Kuzi;ChengXiang Zhai
通讯作者: ChengXiang Zhai