2022 Fiscal Year Research-status Report
Research on analyzing Mongolian legal documents using deep learning
| Project/Area Number |
21K12600
|
|---|
| Research Institution | Ritsumeikan University |
|---|
Principal Investigator |
バトジャルガル ビルゲサイハン 立命館大学, 総合科学技術研究機構, 助教 (30725396)
|
|---|
| Project Period (FY) |
2021-04-01 – 2025-03-31
|
| Keywords | deep learning / legal documents / text classification / Mongolian |
|---|
| Outline of Annual Research Achievements |
In the FY2022, the following tasks have been performed: 1) collecting and preparing training datasets, 2) proposing deep learning models for Mongolian legal documents and, 3) developing a web-based system.
Approximately 4,716 resolutions of the government of Mongolia, 2,066 resolutions of the State Ikh Khural (Parliament of Mongolia), 996 resolutions of self-governing bodies (the Citizens’ Representative Hurals) of the provinces and the capital city Ulaanbaatar, 778 laws of Mongolia, 748 ministerial decrees, 654 International treaties concluded or ratified by Mongolia, 324 regulations of government agencies, 295 decisions of the Constitutional Court of Mongolia, 211 decrees of the president of Mongolia, 174 resolutions of the State Supreme Court, 169 decisions of various councils, committees and other collectives, 114 decisions of the Heads of the bodies appointed by the Parliament, and 78 orders of the governors of the provinces and the mayor of the capital city Ulaanbaatar were prepared.
We also proposed deep-learning-based models for classifying modern Mongolian documents and preliminary experiments were conducted.
Moreover, in the FY2022, based on the research results obtained in the past, we developed a web-based system to classify modern Mongolian documents. Currently, the proposed web-based system shows only the classification results of ordinary modern Mongolian documents. However, in the next FYs, more analysis of Mongolian legal documents will be integrated into the system.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
Our research has been conducted according to the research plan. As planned, the achievements in the FY2022 allowed advancing my research towards to the research goal in developing a deep-learning-based method to analyze Mongolian legal documents. Surveys were conducted in Mongolia, as well as feedbacks and data were obtained through business trips. Ongoing research results and achievements have been published in parts in two international conference papers.
|
| Strategy for Future Research Activity |
In the FY2023, business trips for 1) conducting evaluations among overseas users, and 2) obtaining analyses and feedback from face-to-face meetings will also be conducted. We plan to conduct evaluations at the National University of Mongolia and Ritsumeikan University in Japan. The proposed system will be evaluated by 1) conducting experiments and calculating standard measures such as precision, recall and F-measure; and 2) user evaluations among experts and users who have tried the proposed system. Further improvements to the method will be carried out based on the evaluation results and user feedback.
In the AY2023, the proposed web-based system will be enhanced by adding more analysis of Mongolian legal documents.
|
| Causes of Carryover |
The remaining budget has occurred due to the leftover of the previous year’s such as restrictions and bans of the COVID-19 in FY2020. I will use the remaining budget for next years’ research activities that requires more budgets. These activities include 1) to develop web-based systems and create websites; 2) to hire experts and students on a part-time basis to a) evaluate the proposed system b) conduct experiments, and c) analyze the experimental results; as well as 3) to make business trips to a) meet and obtain feedback, advices, evaluations from the researchers at the National University of Mongolia, and b) present research achievements and results at the domestic and international conferences.
|
