ECHR數據集是有關人權的判例法，目前除了美國、英國和一些國家採用之外，許多國家採用成文法，所以有關判例法的研究較為有限。然而，諸如ECHtR、CVRIA和ILO之類的國際組織不僅使用成文法，而且還運用判例法，代表這樣的大型組織需要開發AI法官系統作為補助來解決瑣碎的案件，因此，關於判例法的研究會很有發展。此外，仍然有一些國家需要保護人權，而AI法官可以給出更客觀的觀點，補助法官有效地閱讀和理解各種先例，從而做出更好的判斷。
在這項研究中，我們的目標是預測歐洲法院在人權案件中的判決結果。我們利用了先前研究提供的ECHR數據集，而且刪除遺漏案例，以進一步完善該資料集。另外，我們提出了Word2Vec加上混合模型，其中Word2Vec提供在輸入階段已含有的單詞之間相關性資訊，而混合模型為結合雙向GRU、卷積神經網絡(CNN)和注意力機制(Attention Mechanism)來進行深度學習的訓練，該混合模型不僅從輸入案件學習循序文本信息的長期依賴性，並且擷取局部特徵，來應對這一個挑戰，最後獲得一個可以判決案件是否侵害人權的分類模型。
在實驗中，我們使用支持向量機器(SVM)為基準和各種深度學習方法來進行比較；根據我們的實驗結果，在深度學習模型中，我們最好的方法是混合模型，其在10摺疊交叉驗證上的準確率達到83％，相較於以往研究的結果，優於他們所獲得的75％，而我們所提的混合模型在外部測試的準確率達到71％，且比較穩定。

The ECHR dataset is a human rights dataset based on Case law. Currently, most countries adopt Written law, except the United States, the United Kingdom, and some countries. Therefore, the research on Case law is far limited. However, we see huge potential for researches on Case law because international organizations such as ECHtR, CVRIA, and ILO not only apply Written law but also Case law, which implies that such a large organization would need to develop an AI judge system as a subsidy to solve trivial cases in the future. It makes research on Case law is promising. Furthermore, there are still countries that need the protection of human rights, which AI judges could give a more objective perspective for assisting the judges to read and understand various precedent cases efficiently, resulting in better judgment.
In this research, our goal is to predict the result in the human rights cases of the European court. We utilize the ECHR dataset provided from the previous research, which we further refine removing missing cases. Besides, we propose the Word2Vec embedding and Hybrid model that combines BiGRU, CNN, and Attention mechanism to address this challenge. After that, we attain a judgment result whether the case is a violation or non-violation. Word2Vec embedding provides a semantical relation in the input level. The proposed hybrid model not only learns the local features but also captures the long-term dependencies of sequential text information.
In the experiment, we use a support vector machine (SVM) and various deep learning methods as a baseline and make comparisons between them; according to our experimental results, our best method is the hybrid model among all the used deep learning models. It achieves an accuracy of 83% from 10-fold cross-validation, which is better than the accuracy of 75% obtained by the previous research on SVM. Moreover, The accuracy of our proposed hybrid model reaches relatively stable 71% on the test dataset.

[1] London, the United Kingdom, equality human rights. (2010). Accessed: Feb. 14, 2020. [Online]. Available: https://www.equalityhumanrights.com/en/what-european-convention-human-rights
[2] Strasbourg Cedex, France, European Court of Human Rights. (1997). Accessed: Feb. 16, 2020. [Online]. Available: https://hudoc.echr.coe.int/eng#{%22documentcollectionid2%22:[%22GRANDCHAMBER%22,%22CHAMBER%22]}
[3] A. E. C. Basave et al., “Automatic labelling of topic models learned from twitter by summarization,” in Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, vol. 2, Jun. 23-25, pp. 618-624, 2014.
[4] P. Mathur et al., “Language Identification from Text Documents,” 2015, arXiv: 1701.03682.
[5] S. Semmler and Z. Rose, “Artificial intelligence : Application today and implications tomorrow,” Duke L. & Tech. Rev., vol. 16, pp. 85-99, 2017.
[6] J. Dressel et al., “The accuracy, fairness, and limits of predicting recidivism,” Science advances, vol. 4, no. 1, 2018. [Online]. Available: https://advances.sciencemag.org/content/4/1/eaao5580.full
[7] V. Vapnik and V. Vapnik, Statistical Learning Theory, New York, USA: Wiley, vol. 1, pp. 624, 1998.
[8] N. Aletras et al., “Predicting judicial decisions of the European Court of Human Rights: a Natural Language Processing perspective,” PeerJ Computer Science, Oct. 24, 2016. [Online].Available: https://peerj.com/articles/cs-93/
[9] M. Medvedeva et al., “Using machine learning to predict decisions of the European Court of Human Rights,” in Artificial Intelligence and Law, Jun. 26, pp. 1-31, 2019. [Online]. Available: https://link.springer.com/content/pdf/10.1007/s10506-019-09255-y.pdf
[10] A. Quemy et al., “European Court of Human Right Open Data project,” 2018, arXiv:1810.03115.
[11] C. O'Sullivan et al., “Predicting the Outcome of Judicial Decisions made by the European Court of Human Rights,” in Proceedings for the 27th AIAI Irish Conference on Artificial Intelligence and Cognitive Science, Galway, Ireland, pp. 272-283, 2019.
[12] OM. Sulea et al., “Exploring the use of text classification in the legal domain,” 2017, arXiv:1710.09306.
[13] I. Chalkidis et al., “Neural Legal Judgment Prediction in English,” 2019, arXiv:1906.02059.
[14] Z. Yang et al., “Hierarchical Attention Networks for Document Classification,” in Proceedings of the 15th Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, San Diego, California, pp. 1480-1489, 2016.
[15] Ching Y. Suen, “n-Gram Statistics for Natural Language Understanding and Text Processing,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. PAMI-1, no. 2, pp. 164-172, 1979.
[16] European Convention on Human Rights, European Court of Human Rights. (1984). Accessed: Jan. 3, 2020. [Online]. Available: https://www.echr.coe.int/Documents/Convention_ENG.pdf
[17] T. Mikolov et al., “Efficient estimation of word representations in vector space,” in Proceedings of the 1st International Conference on Learning Representations, Scottsdale, Arizona, pp. 1-12, 2013.
[18] G. E. Hinton, “Learning distributed representations of concepts,” in Proceedings of the Eighth Annual Conference of the Cognitive Science Society, Amherst, Massachusetts, pp. 1-12, 1986.
[19] J. Chung et al., “Empirical evaluation of gated recurrent neural networks on sequence modeling,” 2014, arXiv:1412.3555.
[20] K. Cho et al., “On the properties of neural machine translation: Encoder-decoder approaches,” in Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation, Doha, Qatar, pp. 103–111, 2014.
[21] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural computation, vol. 9, no. 8, pp. 1735-1780, 1997.
[22] MT. Luong et al., “Effective approaches to attention-based neural machine translation,” in Proceedings of the International Conference on Empirical Methods in Natural Language Processing, Lisbon, Portugal, pp. 1412-1421, 2015.
[23] Y. LeCun et al., “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998.
[24] Y. Kim et al., “Convolutional neural networks for sentence classification,” in Proceedings of the International Conference on Empirical Methods in Natural Language Processing, Doha, Qatar, pp.1746-1751, 2014.
[25] M. Hughes et al., “Medical text classification using convolutional neural networks,” Stud Health Technol Inform, vol. 235, pp. 246-250, 2017.