Citation paper supplement
Browse files
README.md
CHANGED
@@ -20,15 +20,31 @@ license: "apache-2.0"
|
|
20 |
|
21 |
|
22 |
|
23 |
-
# bert-ancient-chinese
|
24 |
|
25 |
-
|
|
|
|
|
26 |
|
27 |
With the current wave of Artificial Intelligence and Digital Humanities sweeping the world, the automatic analysis of modern Chinese has achieved great results. However, the automatic analysis and research of ancient Chinese is relatively weak, and it is difficult to meet the actual needs of Sinology, history, philology, Chinese history and the education of Sinology and traditional culture. There are many controversies about characters, words and parts of speech in ancient Chinese, and there are many difficulties in resource construction. Digital Humanities research requires large-scale corpora and high-performance ancient natural language processing tools. In view of the fact that pre-trained language models have greatly improved the accuracy of text mining in English and modern Chinese texts, there is an urgent need for pre-trained models for the automatic processing of ancient texts.
|
28 |
|
29 |
In 2022, we took part in **[EvaHan 2022](https://circse.github.io/LT4HALA/2022/EvaHan)**, the first NLP tool evaluation competition in the field of ancient Chinese. **`bert-ancient-chinese`** is trained to further optimize the model effect in open environment.
|
30 |
|
31 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
|
33 |
## Further Pre-training
|
34 |
|
@@ -39,6 +55,8 @@ You can view the introduction of the Chinese version through [this link](https:/
|
|
39 |
|
40 |
- Based on the idea of `Domain-Adaptive Pretraining`, `bert-ancient-chinese` was trained on the basis of `bert-base-chinese ` and was combined with ancient Chinese corpus to obtain a pre-trained model for the field of automatic processing of ancient Chinese.
|
41 |
|
|
|
|
|
42 |
## How to use
|
43 |
|
44 |
### Huggingface Transformers
|
@@ -53,6 +71,8 @@ tokenizer = AutoTokenizer.from_pretrained("Jihuai/bert-ancient-chinese")
|
|
53 |
model = AutoModel.from_pretrained("Jihuai/bert-ancient-chinese")
|
54 |
```
|
55 |
|
|
|
|
|
56 |
## Download PTM
|
57 |
|
58 |
The model we provide is the `PyTorch` version.
|
@@ -71,6 +91,8 @@ Download address:
|
|
71 |
| :------------------: | :----------------------------------------------------------: |
|
72 |
| bert-ancient-chinese | [Link](https://pan.baidu.com/s/1JC5_64gLT07wgG2hjzqxjg ) Extraction code: qs7x |
|
73 |
|
|
|
|
|
74 |
## Evaluation & Results
|
75 |
|
76 |
We tested and compared different pre-trained models on the training and test sets provided by the competition [EvaHan 2022](https://circse.github.io/LT4HALA/2022/EvaHan). We compare the performance of the models by fine-tuning them on the downstream tasks of `Chinese Word Segmentation(CWS)` and `part-of-speech tagging(POS Tagging)`.
|
@@ -115,20 +137,27 @@ We use `BERT+CRF` as the baseline model to compare the performance of `siku-bert
|
|
115 |
</tr>
|
116 |
</table>
|
117 |
|
|
|
118 |
## Citing
|
119 |
|
120 |
If our content is helpful for your research work, please quote it in the paper.
|
121 |
|
|
|
|
|
122 |
## Disclaim
|
123 |
|
124 |
The experimental results presented in the report only show the performance under a specific data set and hyperparameter combination, and cannot represent the essence of each model. The experimental results may change due to random number seeds and computing equipment. **Users can use the model arbitrarily within the scope of the license, but we are not responsible for the direct or indirect losses caused by using the content of the project.**
|
125 |
|
|
|
|
|
126 |
## Acknowledgment
|
127 |
|
128 |
`bert-ancient-chinese` is based on [bert-base-chinese](https://huggingface.co/bert-base-chinese) to continue training.
|
129 |
|
130 |
Thanks to Prof. [Xipeng Qiu](https://xpqiu.github.io/) and the [Natural Language Processing Laboratory of Fudan University](https://nlp.fudan.edu.cn/).
|
131 |
|
|
|
|
|
132 |
## Contact us
|
133 |
|
134 |
Pengyu Wang:wpyjihuai@gmail.com
|
|
|
20 |
|
21 |
|
22 |
|
23 |
+
# **bert-ancient-chinese**
|
24 |
|
25 |
+
|
26 |
+
|
27 |
+
## **Introduction**
|
28 |
|
29 |
With the current wave of Artificial Intelligence and Digital Humanities sweeping the world, the automatic analysis of modern Chinese has achieved great results. However, the automatic analysis and research of ancient Chinese is relatively weak, and it is difficult to meet the actual needs of Sinology, history, philology, Chinese history and the education of Sinology and traditional culture. There are many controversies about characters, words and parts of speech in ancient Chinese, and there are many difficulties in resource construction. Digital Humanities research requires large-scale corpora and high-performance ancient natural language processing tools. In view of the fact that pre-trained language models have greatly improved the accuracy of text mining in English and modern Chinese texts, there is an urgent need for pre-trained models for the automatic processing of ancient texts.
|
30 |
|
31 |
In 2022, we took part in **[EvaHan 2022](https://circse.github.io/LT4HALA/2022/EvaHan)**, the first NLP tool evaluation competition in the field of ancient Chinese. **`bert-ancient-chinese`** is trained to further optimize the model effect in open environment.
|
32 |
|
33 |
+
**If you want to refer to our work, you can refer to this [paper](https://aclanthology.org/2022.lt4hala-1.25/):**
|
34 |
+
|
35 |
+
```
|
36 |
+
@inproceedings{wang2022uncertainty,
|
37 |
+
title={The Uncertainty-based Retrieval Framework for Ancient Chinese CWS and POS},
|
38 |
+
author={Wang, Pengyu and Ren, Zhichen},
|
39 |
+
booktitle={Proceedings of the Second Workshop on Language Technologies for Historical and Ancient Languages},
|
40 |
+
pages={164--168},
|
41 |
+
year={2022}
|
42 |
+
}
|
43 |
+
```
|
44 |
+
|
45 |
+
You can view the introduction of the **Chinese version** through [this link](https://github.com/Jihuai-wpy/bert-ancient-chinese).
|
46 |
+
|
47 |
+
|
48 |
|
49 |
## Further Pre-training
|
50 |
|
|
|
55 |
|
56 |
- Based on the idea of `Domain-Adaptive Pretraining`, `bert-ancient-chinese` was trained on the basis of `bert-base-chinese ` and was combined with ancient Chinese corpus to obtain a pre-trained model for the field of automatic processing of ancient Chinese.
|
57 |
|
58 |
+
|
59 |
+
|
60 |
## How to use
|
61 |
|
62 |
### Huggingface Transformers
|
|
|
71 |
model = AutoModel.from_pretrained("Jihuai/bert-ancient-chinese")
|
72 |
```
|
73 |
|
74 |
+
|
75 |
+
|
76 |
## Download PTM
|
77 |
|
78 |
The model we provide is the `PyTorch` version.
|
|
|
91 |
| :------------------: | :----------------------------------------------------------: |
|
92 |
| bert-ancient-chinese | [Link](https://pan.baidu.com/s/1JC5_64gLT07wgG2hjzqxjg ) Extraction code: qs7x |
|
93 |
|
94 |
+
|
95 |
+
|
96 |
## Evaluation & Results
|
97 |
|
98 |
We tested and compared different pre-trained models on the training and test sets provided by the competition [EvaHan 2022](https://circse.github.io/LT4HALA/2022/EvaHan). We compare the performance of the models by fine-tuning them on the downstream tasks of `Chinese Word Segmentation(CWS)` and `part-of-speech tagging(POS Tagging)`.
|
|
|
137 |
</tr>
|
138 |
</table>
|
139 |
|
140 |
+
|
141 |
## Citing
|
142 |
|
143 |
If our content is helpful for your research work, please quote it in the paper.
|
144 |
|
145 |
+
|
146 |
+
|
147 |
## Disclaim
|
148 |
|
149 |
The experimental results presented in the report only show the performance under a specific data set and hyperparameter combination, and cannot represent the essence of each model. The experimental results may change due to random number seeds and computing equipment. **Users can use the model arbitrarily within the scope of the license, but we are not responsible for the direct or indirect losses caused by using the content of the project.**
|
150 |
|
151 |
+
|
152 |
+
|
153 |
## Acknowledgment
|
154 |
|
155 |
`bert-ancient-chinese` is based on [bert-base-chinese](https://huggingface.co/bert-base-chinese) to continue training.
|
156 |
|
157 |
Thanks to Prof. [Xipeng Qiu](https://xpqiu.github.io/) and the [Natural Language Processing Laboratory of Fudan University](https://nlp.fudan.edu.cn/).
|
158 |
|
159 |
+
|
160 |
+
|
161 |
## Contact us
|
162 |
|
163 |
Pengyu Wang:wpyjihuai@gmail.com
|