Huidan Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Huidan Liu has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Hongyu Lin
Chinese Information Processing Laboratory
1 shared publication - Huadong Li
College of Foreign Languages, Shanghai Maritime University, Shanghai, 201306, China
1 shared publication - Le Sun
Nanjing University of Information Science and Technology
1 shared publication Lihua Liu
Shanghai Maritime University, China
1 shared publicationNicholas Jing Yuan
1 shared publication- Wenkai Zhang
Chinese Academy of Sciences, Aerospace Information Research Institute, Beijing, China
1 shared publication
Latest Publications
- Open Access
- 1 October 2024
- Published by SAGE Publications in Sage Open
- Vol. 14 (4)
Multimodal Discourse Studies in the International Academic Community (1997–2023): A Bibliometric Analysis
In today’s world of multimedia communication, the use of multiple modes of discourse is prevalent in various fields. The international academic community has taken an interest in studying multimodality from different perspectives. This paper uses CiteSpace 6.1.R6 to visually analyze literature on multimodal discourse studies (MDS) in the Web of Science (WoS) Core Collection from 1997 to 2023. The results show a significant increase in English research papers on MDS over the past two decades, a development made possible through collaborative efforts and knowledge-sharing among scholars from different regions and institutions. MDS conducted by scholars from across the world reveals diverse research categories and highlights hot issues and social practices. Consequently, the discipline has produced significant social value. The analyses of keywords and literature further reveal that MDS mainly explores language education, politics, children’s education, identity construction, and media. Additionally, the most cited literature on MDS plays a fundamental role in guiding scholars across the world to understand the key concepts of multimodality and to conduct multimodal discourse analysis from different perspectives. This research provides valuable insights for scholars engaged in or interested in multimodal discourse analysis worldwide, helping them understand MDS’s evolution while offering significant implications for conducting research on multimodal discourse.
- 1 October 2024
- Published by SAGE Publications in Sage Open
- Vol. 14 (4)
Open AccessMultimodal Discourse Studies in the International Academic Community (1997–2023): A Bibliometric Analysis
In today’s world of multimedia communication, the use of multiple modes of discourse is prevalent in various fields. The international academic community has taken an interest in studying multimodality from different perspectives. This paper uses CiteSpace 6.1.R6 to visually analyze literature on multimodal discourse studies (MDS) in the Web of Science (WoS) Core Collection from 1997 to 2023. The results show a significant increase in English research papers on MDS over the past two decades, a development made possible through collaborative efforts and knowledge-sharing among scholars from different regions and institutions. MDS conducted by scholars from across the world reveals diverse research categories and highlights hot issues and social practices. Consequently, the discipline has produced significant social value. The analyses of keywords and literature further reveal that MDS mainly explores language education, politics, children’s education, identity construction, and media. Additionally, the most cited literature on MDS plays a fundamental role in guiding scholars across the world to understand the key concepts of multimodality and to conduct multimodal discourse analysis from different perspectives. This research provides valuable insights for scholars engaged in or interested in multimodal discourse analysis worldwide, helping them understand MDS’s evolution while offering significant implications for conducting research on multimodal discourse.
- Open Access
Denoising Distantly Supervised Named Entity Recognition via a Hypergeometric Probabilistic Model
Denoising is the essential step for distant supervision based named entity recognition. Previous denoising methods are mostly based on instance-level confidence statistics, which ignore the variety of the underlying noise distribution on different datasets and entity types. This makes them difficult to be adapted to high noise rate settings. In this paper, we propose Hypergeometric Learning (HGL), a denoising algorithm for distantly supervised NER that takes both noise distribution and instance-level confidence into consideration. Specifically, during neural network training, we naturally model the noise samples in each batch following a hypergeometric distribution parameterized by the noise-rate. Then each instance in the batch is regarded as either correct or noisy one according to its label confidence derived from previous training step, as well as the noise distribution in this sampled batch. Experiments show that HGL can effectively denoise the weakly-labeled data retrieved from distant supervision, and therefore results in significant improvements on the trained models.
Open AccessDenoising Distantly Supervised Named Entity Recognition via a Hypergeometric Probabilistic Model
Denoising is the essential step for distant supervision based named entity recognition. Previous denoising methods are mostly based on instance-level confidence statistics, which ignore the variety of the underlying noise distribution on different datasets and entity types. This makes them difficult to be adapted to high noise rate settings. In this paper, we propose Hypergeometric Learning (HGL), a denoising algorithm for distantly supervised NER that takes both noise distribution and instance-level confidence into consideration. Specifically, during neural network training, we naturally model the noise samples in each batch following a hypergeometric distribution parameterized by the noise-rate. Then each instance in the batch is regarded as either correct or noisy one according to its label confidence derived from previous training step, as well as the noise distribution in this sampled batch. Experiments show that HGL can effectively denoise the weakly-labeled data retrieved from distant supervision, and therefore results in significant improvements on the trained models.