Inteligência artificial e direito
o impacto transformador na organização de grandes acervos de textos jurídicos
DOI:
https://doi.org/10.35699/2965-6931.2023.47689Palavras-chave:
processamento de linguagem natural, modelos de linguagem, modelagem de tópicos, direito, organização de acervos jurídicosResumo
Os recentes avanços na área de inteligência artificial e de processamento de linguagem natural têm impulsionado diversas mudanças no campo jurídico. No Brasil, o Direito segue um movimento constante de modernização, principalmente direcionado à transparência e ao acesso à informação. O grande volume de documentos jurídicos abre espaço para o desenvolvimento e uso de ferramentas inteligentes que busquem organizar e facilitar a gestão desse acervo. Neste trabalho, mostramos como o uso de modelos de linguagem em conjunto com técnicas de modelagem de tópicos são capazes de organizar e extrair conhecimento desses grandes acervos jurídicos, revelando temas que muitas vezes são implícitos e desconhecidos, o que traz benefícios para diversas aplicações, como a pesquisa por documentos similares e a recomendação de textos jurídicos.
Downloads
Referências
ARORA, S. et al. Contextual embeddings: When are they worth it? arXiv preprint arXiv:2005.09117 (2020).
BARTHOLOMEW, J. et al. How the media is covering ChatGPT. Disponível em: <https://www.cjr.org/tow_center/media-coverage-chatgpt.php> Acesso em 12 ago. 2023
BIANCHI, Federico; TERRAGNI, Silvia; HOVY, Dirk. Pre-training is a hot topic: Contextualized document embeddings improve topic coheren ce. In: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers). [S.l.]: Association for Computational Linguistics, 2021. p. 759–766.
BIEVER, C. ChatGPT broke the Turing test — the race is on for new ways to assess AI. Disponível em: <https://www.nature.com/articles/d41586-023-02361-7> Acesso em 12 ago. 2023
BRASIL. Conselho Nacional de Justiça. Painel de projetos com inteligência artificial no Poder Judiciário, 2020. Disponível em: <https://paineisanalytics.cnj.jus.br/single/?appid=29d710f7-8d8f-47be-8af8-a9152545b771&sheet=b8267e5a-1f1f-41a7-90ff-d7a2f4ed34ea&lang=pt-BR&opt=ctxmenu,currsel>. Acesso em: 14 ago. 2023.
BRASIL. Tribunal Eleitoral Regional do Espírito Santo. Bel, a assistente virtual do TRE-ES, vence a categoria "Inovação Tecnológica" do Prêmio de Inovação Judiciário Exponencial. Comunicação, 06 out. 2021. 2021a. Disponível em: <https://www.tre-es.jus.br/comunicacao/noticias/2021/Outubro/bel-a-assistente-virtual-do-tre-es-vence-a-categoria-inovacao-tecnologica-do-premio-de-inovacao-judiciario-exponencial>. Acesso em: 14 out. 2023.
BRASIL, Tribunal de Justiça do Amazonas. TJAM automatiza classificação de petições intermediárias no Portal e-SAJ. Imprensa, 19 dez. 2019. Disponível em: <https://www.tjam.jus.br/index.php/menu/sala-de-imprensa/2387-tjam-automatiza-classificacao-de-peticoes-intermediarias-no-portal-e-saj>. Acesso em: 14 ago. 2023.
BRASIL. Tribunal de Justiça do Distrito Federal e Territórios. TJDFT lidera número de projetos de Inteligência Artificial no Poder Judiciário. Janeiro de 2021b. Disponível em: <https://www.tjdft.jus.br/institucional/imprensa/noticias/2021/janeiro/tjdft-e-o-tribunal-com-mais-projetos-de-inteligencia-artificial>. Acesso em: 14 ago. 2023.
BRASIL. Supremo Tribunal Federal. Ministra Rosa Weber lança robô VitórIA para agrupamento e classificação de processos, 17 mai. 2023. Disponível em: https://portal.stf.jus.br/noticias/verNoticiaDetalhe.asp?idConteudo=507426&ori=1. Acesso em: 14 ago. 2023.
BROWN, T. B. et al. Language models are few-shot learners. In: Proceedings of the 34th International Conference on Neural Information Processing Systems. Red Hook, NY, USA: Curran Associates Inc., 2020. (NIPS’20). ISBN 9781713829546.
CHALKIDIS, I. et al. Extreme multi-label legal text classification: A case study in EU legislation. In: Proceedings of the Natural Legal Language Processing Workshop 2019. Minneapolis, Minnesota: Association for Computational Linguistics, 2019. p. 78–87. Disponível em: <https://aclanthology.org/W19-2209>.
CHEN, H. et al. A comparative study of automated legal text classification using random forests and deep learning. Information Processing Management, v. 59, n. 2, p. 102798, 2022. ISSN 0306-4573. Disponível em: <https://www.sciencedirect.com/science/article/ pii/S0306457321002764>.
DEVLIN, J. et al. BERT: Pre-training of deep bidirectional transformers for language understanding. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). [S.l.: s.n.], 2019. p. 4171–4186.
DEVLIN, J. et al. BERT: Pre-training of deep bidirectional transformers for language understanding. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). Minneapolis, Minnesota: Association for Computational Linguistics, 2019. p. 4171–4186. Disponível em: <https://aclanthology.org/N19-1423>.
DIENG, Adji. B.; RUIZ, Francisco J. R.; BLEI, David M. Topic modeling in embedding spaces. Transactions of the Association for Computational Linguistics, MIT Press, Cambridge, MA, v. 8, p. 439–453, 2020.
ETHAYARAJH, K. BERT, ELMo, & GPT-2: How Contextual are Contextualized Word Representations? Disponível em: <http://ai.stanford.edu/blog/contextual/> Acesso em 12 ago. 2023
GROOTENDORST, Maarten. Bertopic: Neural topic modeling with a class-based tf-idf procedure. CoRR, abs/2203.05794, 2022. Citado na página 1.
HOFMANN, V. et al. Dynamic contextualized word embeddings. arXiv preprint arXiv:2010.12684 (2020).
HUANG, Z. et al. Context-aware legal citation recommendation using deep learning. In: Proceedings of the Eighteenth International Conference on Artificial Intelligence and Law. New York, NY, USA: Association for Computing Machinery, 2021. (ICAIL ’21), p. 79–88. ISBN 9781450385268. Disponível em: <https://doi.org/10.1145/3462757.3466066>.
JOJRI, P. et al. Natural language processing: History, evolution, application, and future work. In: Proceedings of 3rd International Conference on Computing Informatics and Networks: ICCIN 2020, pp. 365-375. Springer Singapore, 2021.
LE, Quoc; MIKOLOV Tomas. Distributed representations of sentences and documents. In: Proceedings of the 31st International Conference on International Conference on Machine Learning - ICML. [S.l.: s.n.], 2014. p. II–1188–II–1196.
LIU, Q. et al. A survey on contextual embeddings. arXiv preprint arXiv:2003.07278, 2020
LUND, K.; BURGESS C. Producing high-dimensional semantic spaces from lexical co-occurrence. In: Behavior research methods, instruments, & computers 28.2 (1996): 203-208. Disponível em: <https://doi.org/10.3758/BF03204766>
MIKOLOV, T. et al. Efficient Estimation of Word Representations in Vector Space. Proceedings of the First International Conference on Learning Representations, 2013.
RAHMAN, M. F. et al. Hdbscan: Density based clustering over location based services. ArXiv, abs/1602.03730, 2016.
RUSSEL, S. J. Artificial intelligence a modern approach. 3. ed. Pearson Education, Inc., 2010.
SALOMÃO, Luis Felipe (coord.). Tecnologia aplicada à gestão dos conflitos no âmbito do Poder Judiciário Brasileiro. Rio de Janeiro: Editora FGV Conhecimento, 2023.
SANSONE, Carlo; SPERLÍ, Giancarlo. Legal information retrieval systems: State-of-the-art and open issues. Inf. Syst., Elsevier Science Ltd., GBR, v. 106, n. C, may 2022. ISSN 0306-4379. Disponível em: <https://doi.org/10.1016/j.is.2021.101967>.
SCHWARCZ D. et al. AI Tools for Lawyers: A Practical Guide. In: Minnesota Law Review Headnotes. Disponível em: <http://dx.doi.org/10.2139/ssrn.4404017>
SHAO, Y. et al. Bert-pli: Modeling paragraph-level interactions for legal case retrieval. In: Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence. [S.l.: s.n.], 2021. (IJCAI’20). ISBN 9780999241165.
SRIVASTAVA, Akash; SUTTON, Charles. Autoencoding Variational Inference for Topic Models. arXiv e-prints, p. arXiv:1703.01488, mar. 2017.
SULIS, E. et al. Exploiting co-occurrence networks for classification of implicit inter- relationships in legal texts. Information Systems, v. 106, p. 101821, 2022. ISSN 0306-4379. Disponível em: <https://www.sciencedirect.com/science/article/pii/S0306437921000648>.
VASWANI, A. et al. Attention is all you need. In: GUYON, I. et al. (Ed.). Advances in Neural Information Processing Systems. Curran Associates, Inc., 2017. v. 30. Disponível em: <https://proceedings.neurips.cc/paper_files/paper/2017/file/ 3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf>.
WALDRON, Jeremy. The Concept and the Rule of Law. In: No. 08–35; Public Law & Legal Theory Research Paper Series, Issue November, 2008.
WIEDMANN, G. et al. Does BERT make any sense? Interpretable word sense disambiguation with contextualized embeddings. arXiv preprint arXiv:1909.10430, 2019
YANG, J. et al. Legalgnn: Legal information enhanced graph neural network for recommendation. ACM Trans. Inf. Syst., Association for Computing Machinery, New York, NY, USA, v. 40, n. 2, sep 2021. ISSN 1046-8188. Disponível em: <https://doi.org/10.1145/3469887>.
