AI法官之研究：深度學習於刑事訴訟裁判的應用

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：30

、訪客IP：3.149.29.209

姓名

陳勇行(Yung-Sing Chen) 查詢紙本館藏

畢業系所

資訊管理學系在職專班

論文名稱

AI法官之研究：深度學習於刑事訴訟裁判的應用
(Research on AI Judges: Application of Deep Learning in Criminal Trial Adjudication)

相關論文

★ 台灣50走勢分析：以多重長短期記憶模型架構為基礎之預測	★ 以多重遞迴歸神經網路模型為基礎之黃金價格預測分析
★ 增量學習用於工業4.0瑕疵檢測	★ 遞回歸神經網路於電腦零組件銷售價格預測之研究
★ 長短期記憶神經網路於釣魚網站預測之研究	★ 基於深度學習辨識跳頻信號之研究
★ Opinion Leader Discovery in Dynamic Social Networks	★ 深度學習模型於工業4.0之機台虛擬量測應用
★ A Novel NMF-Based Movie Recommendation with Time Decay	★ 以類別為基礎sequence-to-sequence模型之POI旅遊行程推薦
★ A DQN-Based Reinforcement Learning Model for Neural Network Architecture Search	★ Neural Network Architecture Optimization Based on Virtual Reward Reinforcement Learning
★ 生成式對抗網路架構搜尋	★ 以漸進式基因演算法實現神經網路架構搜尋最佳化
★ Enhanced Model Agnostic Meta Learning with Meta Gradient Memory	★ 遞迴類神經網路結合先期工業廢水指標之股價預測研究

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2029-1-1以後開放)

摘要(中)

此近幾年運用機器學習（Machine learning ）與深度學習（Deep learning）方式對於繁體中文、簡體中文、英文等法律文本進行法律專業術語的標記、法律罪章名稱的分類以及法律量刑範圍的預測逐漸為人們所關注，由此國內因國民法官的議題而也有了AI法官議題的延伸，希冀透過更理性、客觀與一制性的邏輯處理能力準確一致的進行案件的裁量。其中據OpenAI實驗室的測試Generative Pre-trained Transformer (GPT)模型如果參加美國紐約州（State of New York）所舉辦的律師司法考試，成績將可落在平均值前10%的高水準，具有相當優秀的法律知識成果，見其Transformer模型於司法文書的應用上的成果。將中文法律領域於機器學習與深度學習的應用，最先碰到的是法律資料結構化的問題，中華民國法律在資料結構化方面處理不易。由於法律資料的數量龐大且法條種類繁多，資料之間的關係和層次復雜，因此需要進行有效的資料結構化和組織化處理。本次AI法官模型研究中的模型架構分為資料前處理、AI法官系統裁量特徵選取、模型訓練與預測三個階段，資料前處理階段會自司法院資料開放平臺下載全國各級法院1996年1月份至2022年8月份裁判書資料JSON開放資料格式檔案。以Bidirectional Encoder Representati-ons from Transformers進行裁判書文本的處理，可通過對法律裁判書的文本來進行訓練，學習到裁判書中的語意信息和詞彙的關係，並將裁判書內容轉換為向量表示，取出資料集中酒後駕車案件類型的特徵。最後以多任務學習multi-task learning結合自動調整loss weight的方式進行模型的訓練與預測，以同時得到罪責及刑期兩種預測結果，準確率可達到0.95。

摘要(英)

In recent years, the use of machine learning and deep learning methods for tagging legal terminology, classifying legal chapter names, and predicting legal sentencing ranges in legal texts written in Traditional Chinese, Simplified Chinese, and Englis- h has gradually become a topic of interest. This has led to the extension of the AI judge issue in domestic circles due to the issue of citizen judges. The hope is to us- e more rational, objective, and standardized logic processing capabilities to make a- ccurate and consistent judgments in legal cases.According to tests conducted by O- penAI Laboratory, the Generative Pre-trained Transformer (GPT) model can achie- ve high-level scores, placing it in the top 10% on the New York State Bar Exam. This indicates that the model has achieved excellent legal knowledge results, dem- onstrating its effectiveness in the application of transformer models in legal doc- uments.In the application of machine learning and deep learning in the Chinese le- gal field, the first challenge encountered is the problem of legal data structuring. The legal system in the Republic of China (Taiwan) is not easy to handle in terms of data structuring. Due to the vast amount of legal data and the complexity of the relationships and hierarchies between different types of legal provisions, effective data structuring and organization is necessary. In this study of the AI judge model, the model architecture is divided into three stages: data preprocessing, feature sele- ction for the AI judge system′s discretion, and model training and prediction. In the data preprocessing stage, the national court′s judgment data in JSON open data format from January 1996 to August 2022 is downloaded from the Judicial Depart- met Open Data Platform. The Bidirectional Encoder Representations from Trans- formers (BERT) is used to process the text of the judgments. By training on the text of legal judgments, the model learns the semantic information and vocabulary relationships within the judgments and converts the judgment contents into vector representations. Relevant features are extracted from the dataset for cases involve- ng drunk driving. Finally, the model is trained and predicted using multi-task lear- ning and automatic loss weight adjustment to obtain both the prediction results for the criminal liability and the sentence length, with an accuracy rate of 0.95.

關鍵字(中)

★ 深度學習
★ 法律文本
★ 法律專業術語
★ 法律罪章名稱分類
★ 法律量刑範圍預測
★ AI 法官
★ 法律資料結構
★ AI 法官系統
★ 裁量特徵選取

關鍵字(英)

論文目次

第 1 章緒論 ........................................................................................................1
1-1 研究背景 .......................................................................................................1
1-2 研究動機與的.................................................................................................6
1-3 研究貢獻 .......................................................................................................9
1-4 論文流程架構.................................................................................................10
第 2 章文獻探討 .................................................................................................12
2-1 機器學習及深度學習於法律文本系統應用........................................................12
2-2 法律文本特徵辨技術......................................................................................19
第 3 章研究方法 .................................................................................................24
3-1 模型架構 ......................................................................................................24 3-2 刑事案件裁判書資料前處理 ..........................................................................25
3-3 AI 法官系統裁量酒後駕車裁判書特徵分類 ......................................................29
3-4 AI 法官系統多任務學習模型訓練 ...................................................................35
第 4 章實驗結果 ................................................................................................39
4-1 實驗環境與資料集.........................................................................................39 4-2 評估方法......................................................................................................41
4-3 模型準確度 ACC 比較...................................................................................43 4-4 模型效能驗證...............................................................................................44 4-5 參數設定對模型影響.....................................................................................45
4-6 實驗總結......................................................................................................63 4-7 實驗實證......................................................................................................65
第 5 章結論 .......................................................................................................68
5-1 研究總結.......................................................................................................68 5-2 研究限制......................................................................................................69
5-3 未來研究方向...............................................................................................70
參考文獻 ............................................................................................................71

參考文獻

Process of Legal Document
[1] W. Lin, T. Kuo, T. Chang, C. Yen, C. Chen and S. Lin,“Exploiting Machine Learning Models for Chinese Legal Documents Labeling, Case Classification, and Sentencing Prediction”, Computational Linguistics and Chinese Language Processing, Vol. 17, No. 4, pp. 49-68, December 2012
[2] Ｈ. Sun,”基於深度學習之法律關鍵詞擷取”,國立中正大學學位論文,2017
[3] 黃詩淳,邵軒磊,“運用機器學習預測法院裁判──法資訊學之實踐”,月旦法學雜誌, Vol. 270, pp. 86-96,2017
[4] 陳冠群,“中文裁判書之要旨擷取：以最高法院裁判書為例”,國立政治大學學位論文,2018
[5] 黃詩淳,邵軒磊,” 酌定子女親權之重要因素：以決策樹方法分析相關裁判”,國立臺灣大學法學論叢, Vol. 47, No. 1, pp. 299-344,2018
[6] 蘇南,”論人工智慧運用於律師服務的未來展望”,Taiwan Bar Journal, Vol. 22, No. 6, pp. 31-40,2018
[7] 黃詩淳,邵軒磊,”人工智慧與法律資料分析之方法與應用：以單獨親權酌定裁判的預測模型為例”,國立臺灣大學法學論叢, Vol. 48, No. 4, pp. 2023-2073,2019
[8] 邵軒磊,吳國清,“法律資料分析與文字探勘：跨境毒品流動要素與結構研究”,問題與研究, Vol. 48, No. 4, pp. 91-114,2019
[9] 駱浩楠,汪峥,李峰,“AppIication of Text SimiIarity Comparison Based on Machine Learning: A Case Study on Text of Law”, Industrial Control Computer, Vol. 33, No. 6, pp. 3-5,2020
[10] 吳志勇,“人工智慧於司法資料系統之應用”,國立中興大學學位論文,2019
[11] 曾韵執等人,“運用自動化與人工智慧提升法令遵循風險管理效率”,勤業眾信,2018
[12] Lawsnote Monta, https://monta.lawsnote.com/landing-page
[13] 黃詩淳,邵軒磊,“以人工智慧讀取親權酌定裁判文本：自然語言與文字探勘之實踐”, 國立臺灣大學法學論叢, Vol. 49, No. 1, pp. 195-224,2020
[14] 李榮耕,“刑事程序中人工智慧於風險評估上的應用”,政大法學評論, Vol. 168, pp. 117-186,2022
[15] 顧以謙,張道行,許福元,吳瑜,林俐如,宋曜廷,李思賢,“應用 AI 人工智慧自動判讀起訴書類先導研究－以施用毒品罪為例”, 刑事政策與犯罪防治研究專刊,Vol. 30, pp. 93-140,2021
[16] 林常青,“Judges′ Sentencing Decisions, Anchoring Effects, and Sentence Guideline Information System in Taiwan”,國立成功大學經濟學系科技部研究計畫,2019
[17] 鍾文傑,陳哲文,王駿發,曾世邦,王宗松,“基於多 BERT 模型之 NLLP 應用於建築工程訴訟之理解與預測”, ROCLING論文集,2020
[18] 鄭明政,“從State v. Loomis 案件看AI 應用於司法審判上的若干問題”,台日法政研究, Vol. 4, pp. 165-178,2020
[19] 邵軒磊,“新住民相關親權酌定裁判書的文字探勘：對「平等」問題的法實證研究嘗試”, 國立臺灣大學法學論叢, Vol. 49, No. S, pp. 1267-1308,2020
[20] Y. Danding,“Case Study of Criminal Law Based on Multi-task Learning”, 2020 International Conference on Artificial Intelligence and Computer Engineering (ICAICE), pp. 98-103,2020
[21] 邵軒磊,“人工智慧與酒駕刑度估計──深度學習卷積神經網路量刑模型之實踐”,月旦法學雜誌, Vol. 312, pp. 105-116,2021
[22] 曹錫璋,“基於深度學習模型之判決書情境相似檢索技術之研究”,國立中興大學學位論文,2021
[23] 江沂璇,“論判決預測系統對親權酌定案件之衝擊”,國立清華大學學位論文,2020
[24] Bert of Legal Document
[25] J. Devlin, M. Chang, K. Lee, K. Toutanova,“BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding”,arXiv,2017
[26] Nut Limsopatham,“Effectively Leveraging BERT for Legal Document Classification”, Proceedings of the Natural Legal Language Processing Workshop,2021
[27] F. Aman, W. Yanchuan,“Multi task intelligent legal judgment method based on Bert model”, Microelectronics & Computer, Vol. 39, No. 9, pp. 107-114,2022
[28] S. Monika, B. Simone, C. Giulia and S. Kuno,“Smart criminal justice: exploring the use of algorithms in the Swiss criminal justice system”, Artificial Intelligence and Law, Vol. 31, No. 2, pp. 213-237,2022
[29] Z. Min, L. Bo, S. Le, K. Zhao, Z. Kaifa,“Study of Deep Learning-Based Legal Judgment Prediction in Internet of Things Era”, Computational intelligence and neuroscience, Vol. 2022, pp. 1-6,2022
[30] W. Yongjun, G. Jing, C. Junjie,“Deep Learning Algorithm for Judicial Judgment Prediction Based on BERT”, 2020 5th International Conference on Computing, Communication and Security (ICCCS), Vol. 2020, pp. 1-6,2020
[31] Transformer of Legal Document
[32] N. Ha-Thanh, N. Minh-Phuong, V. Thi-Hai-Yen, B. Minh-Quan, N. Minh-Chau, D. Tran-Binh, T. Vu, N. Le-Minh and S. Ken,“Transformer-based Approaches for Legal Text Processing”, arXiv,2022
[33] Multi-task Learning of Legal Document
[34] C. Roberto, G. Yarin, K. Alex,“Multi-task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics”, IEEE/CVF Confere-nce on Computer Vision and Pattern Recognition, pp. 7482-7491,2022
[35] R. Sebastian,“An Overview of Multi-Task Learning in Deep Neural Networks”,arXiv,2017
[36] CNN of Legal Document
[37] 莊東穎,陳裕賢,吳孟倫,“基於人工智慧之刑法量刑預測技術和系統”, 國立臺北大學資訊工程學系科技部研究計畫,2020
[38] 吳孟倫,“基於人工智慧的法律科技技術研發–法學資料檢索與案件結果預測”,私立淡江大學資訊工程學系科技部研究計畫,2021
[39] H. Decheng, C. Lieuhen, S. Taiping,“Legal judgment prediction based on machine learning: Predicting the discretionary damages of mental sufferi-ng in fatal car accident cases”,Applied sciences, Vol. 11, No. 21, pp. 10361,2021
[40] H. Qian, K. Yufeng, S. Derek,“Experimental Evaluation of CNN Param-eters for Text Categorization in Legal Document Review”,2019 IEEE International Conference on Big Data (Big Data), Vol. 2019, pp. 4320-4324,2019
[41] C. Rishi, K. Robert, G. Peter, H. Nathaniel, Z. Jianping, Z. Haozhen,“CNN Application in Detection of Privileged Documents in Legal Document Review”, 2020 IEEE International Conference on Big Data (Big Data), Vol. 2020, pp.1485-1492,2020
[42] K. Robert, C. Rishi, H. Nathaniel, Z. Jianping, W. Fusheng, Z. Haozhen, S. Ye, Q. Han,“2019 IEEE International Conference on Big Data (Big Data)”, Vol. 2019, pp.2038-2042,2019
[43] H. Qian, S. Derek,“Comparison of Deep Learning Technologies in Legal Document Classification”, 2021 IEEE International Conference on Big Data (Big Data), Vol. 2021, pp.2701-2704,2021
[44] LSTM of Legal Document
[45] G. Osvaldo, M. Beniamino, M. Sanjay, G. Chiara, B. Ivan, T. David, A. Bernady, R. Ana, T. Eufemia, T. Carmelo,“A LSTM Recurrent Neural Network Implementation for Classifying Entities on Brazilian Legal Documents”, Computational Science and Its Applications - ICCSA 2021, Vol.12950, pp.648-656,2021
[46] N. Qadri, Y. Evi,“Legal Entity Recognition in Indonesian Court Decision Documents Using Bi-LSTM and CRF Approaches”, 2020 International Conference on Advanced Computer Science and Information Systems (ICACSIS), , pp.429-434,2020
[47] L. Guodong, W. Zhe, Ma, Yinglong,“Combining Domain Knowledge Extraction With Graph Long Short-Term Memory for Learning Classification of Chinese Legal Documents”, IEEE access, Vol.7, pp.139616-139627, 2019

指導教授

陳以錚(Yi-Cheng Chen)

審核日期

2024-1-19

推文