Chengwen Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Chengwen Wang has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Dianting Liu
Guilin University of Technology
1 shared publication Endong Xun
1 shared publicationMiao Zhou
1 shared publicationQingxiu Dong
MOE Key Lab of Computational Linguistics, School of Computer Science, Peking University, Beijing, China
1 shared publicationXiaochen Wang
MOE Key Lab of Computational Linguistics, School of Computer Science, Peking University, Beijing, China
1 shared publication- Xiaodan Zhou
North Carolina State University Department of Statistics Raleigh, NC 27695
1 shared publication
Latest Publications
- Open Access
- 6 September 2024
- Published by Cambridge University Press (CUP) in Natural Language Processing
- Vol. 31 (1) , 90-110
Statistical dataset evaluation: A case study on named entity recognition
Datasets serve as crucial training resources and model performance trackers. However, existing datasets have exposed a plethora of problems, inducing biased models and unreliable evaluation results. In this paper, we propose a model-agnostic dataset evaluation framework for automatic dataset quality evaluation. We seek the statistical properties of the datasets and address three fundamental dimensions: reliability, difficulty, and validity, following a Classical Test Theory (CTT). Taking the named entity recognition (NER) datasets as a case study, we introduce nine statistical metrics for a statistical dataset evaluation framework. Specifically, we investigate the reliability of a NER dataset with three metrics, including Redundancy, Accuracy, and Leakage Ratio. We assess the dataset difficulty through four metrics: Unseen Entity Ratio, Entity Ambiguity Degree, Entity Density, and Model Differentiation. For validity, we introduce the Entity Imbalance Degree and Entity-Null Rate to evaluate the effectiveness of the dataset in assessing language model performance. Experimental results validate that our evaluation framework effectively assesses various aspects of the dataset quality. Furthermore, we study how the dataset scores on our statistical metrics affect the model performance and appeal for dataset quality evaluation or targeted dataset improvement before training or testing models.
- 6 September 2024
- Published by Cambridge University Press (CUP) in Natural Language Processing
- Vol. 31 (1) , 90-110
Open AccessStatistical dataset evaluation: A case study on named entity recognition
Datasets serve as crucial training resources and model performance trackers. However, existing datasets have exposed a plethora of problems, inducing biased models and unreliable evaluation results. In this paper, we propose a model-agnostic dataset evaluation framework for automatic dataset quality evaluation. We seek the statistical properties of the datasets and address three fundamental dimensions: reliability, difficulty, and validity, following a Classical Test Theory (CTT). Taking the named entity recognition (NER) datasets as a case study, we introduce nine statistical metrics for a statistical dataset evaluation framework. Specifically, we investigate the reliability of a NER dataset with three metrics, including Redundancy, Accuracy, and Leakage Ratio. We assess the dataset difficulty through four metrics: Unseen Entity Ratio, Entity Ambiguity Degree, Entity Density, and Model Differentiation. For validity, we introduce the Entity Imbalance Degree and Entity-Null Rate to evaluate the effectiveness of the dataset in assessing language model performance. Experimental results validate that our evaluation framework effectively assesses various aspects of the dataset quality. Furthermore, we study how the dataset scores on our statistical metrics affect the model performance and appeal for dataset quality evaluation or targeted dataset improvement before training or testing models.
- 13 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 966-969
A New Paradigm of Computer Art Design based on Fusion of Light Collection Sensors and Smart Image Detection Algorithms
The use of image recognition technology can complete tasks that cannot be achieved by traditional sensor technology, and at the same time, with the help of image recognition and multi-source information fusion, the monitoring effect can be better and more accurate. With the continuous development of computer technology. Based on the principle of sparse representation, integrating cutting-edge computer technology into art design can simplify the difficulty of art design work. This article takes Photoshop technology as an example to illustrate the teaching research of Photoshop in art design work. Select the M nearest neighbors of the test sample to replace the original training sample; weighted decision fusion. The experimental results on the CSIST face database show that, compared with similar algorithms, the proposed algorithm improves the recognition rate and calculation speed.
- 13 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 966-969
A New Paradigm of Computer Art Design based on Fusion of Light Collection Sensors and Smart Image Detection Algorithms
The use of image recognition technology can complete tasks that cannot be achieved by traditional sensor technology, and at the same time, with the help of image recognition and multi-source information fusion, the monitoring effect can be better and more accurate. With the continuous development of computer technology. Based on the principle of sparse representation, integrating cutting-edge computer technology into art design can simplify the difficulty of art design work. This article takes Photoshop technology as an example to illustrate the teaching research of Photoshop in art design work. Select the M nearest neighbors of the test sample to replace the original training sample; weighted decision fusion. The experimental results on the CSIST face database show that, compared with similar algorithms, the proposed algorithm improves the recognition rate and calculation speed.
- Open Access
- 1 September 2008
- Published by Springer Nature in Optoelectronics Letters
- Vol. 4 (5) , 384-386
Wavelet-based multispectral face recognition
- 1 September 2008
- Published by Springer Nature in Optoelectronics Letters
- Vol. 4 (5) , 384-386
Open AccessWavelet-based multispectral face recognition