Using deep learning to predict outcomes of legal appeals better than human experts: A study with data from Brazilian federal courts

Elias Jacob de Menezes-Neto; Marco Bruno Miranda Clementino

doi:10.1371/journal.pone.0272287

Peer Review History

Original SubmissionMarch 18, 2022
4 Jul 2022 Decision Letter - Donrich Thaldar, Editor PONE-D-22-07895Using deep learning to predict outcomes of legal appeals better than human experts: a study with data from Brazilian federal courtsPLOS ONE Dear Dr. Jacob de Menezes-Neto, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please address all the recommendations by the two reviewers. Please submit your revised manuscript by Aug 18 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript: A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Donrich Thaldar Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. In order to improve reporting, in your methods section, please provide additional information about the participant recruitment method and the demographic details of your participants. 3. Thank you for stating the following in the Funding Section of your manuscript: This study was funded by the National Council for Scientific and Technological Development (CNPq) through a scholarship to Elias Jacob de Menezes-Neto (302668/2020-9). The Brazilian Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) financed the fee to publish this article (finance code XXX). Funders had no role in the study design, data collection and analysis, or the decision to prepare and publish the manuscript. We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form. Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows: This study was funded by the National Council for Scientific and Technological Development (CNPq) through a scholarship to EJDMN (302668/2020-9). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Please include your amended statements within your cover letter; we will change the online submission form on your behalf. 4. Thank you for stating the following in the Competing Interests section: I have read the journal's policy and the authors of this manuscript have the following competing interests: Elias Jacob de Menezes-Neto declares no competing interests. Marco Bruno Miranda Clementino is a federal judge in the 5th Regional Federal Court jurisdiction. Although his position is not affected in any way by this paper, we understand that this affiliation may be seen as a non-financial competing interest. Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared. Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf. 5. Please remove your figures from within your manuscript file, leaving only the individual TIFF/EPS image files, uploaded separately. These will be automatically included in the reviewers’ PDF. 6. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information. 7. Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Partly Reviewer #2: Yes ******** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes Reviewer #2: Yes ****** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes Reviewer #2: Yes ****** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes Reviewer #2: Yes ******** 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: * Abstract - Sentence 3: This sentence does not motivate the need for the study as this study does not really help understand the steps of case flow. - Line 8: machines --> models - Line 7 - 9: We compare the predictive performance of the models to the predictions of twenty two highly skilled legal experts. - Line 10: ... of 0.3688 compared to 0.1253 from the human experts. - 2nd last sentence - Rewrite the conclusion to be more specific: Our results demonstrate that natural language machine learning techniques is a promising approach for predicting legal outcomes. Introduction Page 3: - Line 3 : admitted --> submitted - First paragraph: "We propose the ..." - Should specifically mention that the class of models used in this paper are deep learning, natural language processing (NLP) models. - Last line: analysis conducted --> predictions determined ... - Second paragraph: "we evaluated ..." - Line 5: delete "we used" - Line 7: replace the first "architecture" --> "model" - Third paragraph: "We discuss how ... " - Line 1: "an algorithm" --> "models" - Sentence 1: Begin the paragraph with: AI-driven systems could have real potential to improve stability, predictability ... Adust the paragraph appropriately. - While the benefits of such systems are described - it must be mentioned that much more work is required before automated systems replace human decision making. Institutional background - This section must be reduced - while it may be usefull background - much of it is not necessary to understand the rest of the paper. It should rather be reduced from 5 pages to 2 or so pages. - I would rather see the space used to give a background to the NLP models used in the study. NLP Models Section - The paper should include a background on the NLP models used in the study. For e.g. LSTM and Transformer models should be described. Methodology - Paragraph Two: - clarify if any pre-processing was done - either no preprocessing, or if any then what did this entail. - Paragraph Three: - did each participant predict each of the 690 appeals? - In what language are the full text written in? - Page 13 - first two paragraphs - would better fit in the introduction that discusses the motivation for the study. - Page 14: - While the labelling process is sensible, it is not fool proof. Where there any processes used to verify the labels? For e.g. sampling some of the decisions? - Page 16: - First paragraph in "Experimental setup" - ".. to all parties in production. " --> "... to all parties." - Experimental Setup: - The training set up of the different models should give the hyper parameters used and explain how these were tuned. - Some more details of the model set up would be useful if they differed from the standard set up discussed in the literature. - Legal expert's analysis - "... 17 highly skilled ..." This number contradicts the numbers given earlier in the paper - where 22 is reported. Please fix. - Page 19: - While we understand the difficulty to obtain human expert decisions, it would have good to evaluate inter-human agreement i.e. if participants were to have evaluated the same cases. Results - The hyperparameters must be clearly specified in the methods. Discussion - Paragraph Two: - Line 2: "example The" --> "examople. The" - Paragraph Four: - Sentence Two is not substantiated: How will the tool help to detect inconsistencies - as the tool does not, as described, cluster cases on similarity. So this entire paragraph is problematic and should be left out or substantiated carefully. Reviewer #2: The paper provides an interesting contribution to the literature. I believe that the paper could improve the explanation of the models that are used and provide the codes so other researchers can follow these steps. It could open an important reseach area applied to judicial sentences and increase the potential for citations of the paper. ******** 6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No ******** [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0272287.r001
Revision 1
13 Jul 2022 Author Response Natal/Brazil, July 13th, 2022 Dear Editor, We appreciate the careful reading of our manuscript. We have prepared a revised version of the paper that takes into account all of the comments made by the academic editor and reviewers. Below, we reply to the comments, specifying the text changes. We look forward to receiving your decision. Sincerely, Elias Jacob de Menezes-Neto (on behalf of the authors) A) JOURNAL REQUIREMENTS 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. Done. 2. In order to improve reporting, in your methods section, please provide additional information about the participant recruitment method and the demographic details of your participants. Done. Please see lines 374-378. Pursuant to the Brazilian National Council of Health Resolution n. 510/2016, article 1, item VII, researchers in humanities and social sciences (including law) are not required to submit their research for REB review when the research does not involve the collection of personal identification of participants or when the researchers do not publish personally identifiable information and the information collected is related to subjects’ professional practice. For this reason, we could not collect demographic details about participants other than the information already in the manuscript. 3. Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. We removed the funding statement from the manuscript and amended the cover letter to update our funding statement. 4. Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf. Done 5. Remove figures from the Manuscript. Done. 6. Include captions for the Supporting Information at the end of the manuscript. Done. 7. Review reference list We had to include several new items on the reference list to accommodate the requests made by the reviewer #1 to provide more details about our models. These were added: Peters ME, Neumann M, Iyyer M, Gardner M, Clark C, Lee K, et al. Deep contextualized word representations. arXiv; 2018. Available: http://arxiv.org/abs/1802.05365 Soam M, Thakur S. Next Word Prediction Using Deep Learning: A Comparative Study. 2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence). 2022. pp. 653–658. doi:10.1109/Confluence52989.2022.9734151 Hochreiter S, Schmidhuber J. Long Short-Term Memory. Neural Computation. 1997;9: 1735–1780. doi:10.1162/neco.1997.9.8.1735 Hochreiter S. The Vanishing Gradient Problem During Learning Recurrent Neural Nets and Problem Solutions. Int J Unc Fuzz Knowl Based Syst. 1998;06: 107–116. doi:10.1142/S0218488598000094 Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, et al. Attention is all you need. Advances in Neural Information Processing Systems. 2017. pp. 5999–6009. Zhou H, Zhang S, Peng J, Zhang S, Li J, Xiong H, et al. Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting. arXiv; 2020. doi:10.48550/ARXIV.2012.07436 Pappagari R, Zelasko P, Villalba J, Carmiel Y, Dehak N. Hierarchical Transformers for Long Document Classification. 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings. 2019; 838–844. doi:10.1109/ASRU46091.2019.9003958 Rajbhandari S, Rasley J, Ruwase O, He Y. Zero: Memory optimizations toward training trillion parameter models. International Conference for High Performance Computing, Networking, Storage and Analysis, SC. 2020;2020-Novem: 1–24. doi:10.1109/SC41405.2020.00024 Snoek J, Larochelle H, Adams RP. Practical Bayesian Optimization of Machine Learning Algorithms. In: Pereira F, Burges CJ, Bottou L, Weinberger KQ, editors. Advances in Neural Information Processing Systems. Curran Associates, Inc.; 2012. Available: https://proceedings.neurips.cc/paper/2012/file/05311655a15b75fab86956663e1819cd-Paper.pdf Biewald L. Experiment Tracking with Weights and Biases. 2020. Available: https://www.wandb.com/ Howard J, Gugger S. Fastai: A Layered API for Deep Learning. Information. 2020;11: 108. doi:10.3390/info11020108 Smith LN. Cyclical Learning Rates for Training Neural Networks. arXiv; 2017. Available: http://arxiv.org/abs/1506.01186 Smith LN, Topin N. Super-convergence: very fast training of neural networks using large learning rates. In: Pham T, editor. Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications. SPIE; 2019. p. 36. doi:10.1117/12.2520589 Kingma DP, Ba J. Adam: A Method for Stochastic Optimization. arXiv; 2017. Available: http://arxiv.org/abs/1412.6980 Agarap AF. Deep Learning using Rectified Linear Units (ReLU). arXiv; 2019. Available: http://arxiv.org/abs/1803.08375 Misra D. Mish: A Self Regularized Non-Monotonic Activation Function. arXiv; 2020. Available: http://arxiv.org/abs/1908.08681 Loshchilov I, Hutter F. Decoupled Weight Decay Regularization. arXiv; 2019. Available: http://arxiv.org/abs/1711.05101 We have also corrected two references: Before: Souza F, Nogueira R, Lotufo R. Portuguese Named Entity Recognition using BERT-CRF. arXiv preprint arXiv:190910649. 2019. After: Souza F, Nogueira R, Lotufo R. {BERT}imbau: pretrained {BERT} models for {B}razilian {P}ortuguese. 9th Brazilian Conference on Intelligent Systems, {BRACIS}, Rio Grande do Sul, Brazil, October 20-23 (to appear). 2020. Before: Lage-Freitas A, Allende-Cid H, Santana O, de Oliveira-Lage L. Predicting Brazilian court decisions. 2019; 1–4. After: Lage-Freitas A, Allende-Cid H, Santana O, Oliveira-Lage L. Predicting Brazilian Court Decisions. PeerJ Computer Science. 2022;8: e904. doi:10.7717/peerj-cs.904 B) REVIEWER 1 1. ABSTRACT - Sentence 3: This sentence does not motivate the need for the study as this study does not really help understand the steps of case flow. We rephrased the sentence. See lines 27-28. 2. Line 8: machines --> models Done. 3. Line 7 - 9: We compare the predictive performance of the models to the predictions of twenty two highly skilled legal experts. Done. 4. Line 10: ... of 0.3688 compared to 0.1253 from the human experts. Done. 5. 2nd last sentence - Rewrite the conclusion to be more specific: Our results demonstrate that natural language machine learning techniques is a promising approach for predicting legal outcomes. Done. Lines 33-35. 6. INTRODUCTION - Page 3: - Line 3 : admitted --> submitted Done. 7. First paragraph: "We propose the ..." Should specifically mention that the class of models used in this paper are deep learning, natural language processing (NLP) models. Done. 8. Last line: analysis conducted --> predictions determined .. Done. 9. Second paragraph: "we evaluated ..." We did not understand what the reviewer meant here. 10. Line 5: delete "we used" Done. 11. Line 7: replace the first "architecture" --> "model" Done. 12. Third paragraph: "We discuss how ... " We did not understand what the reviewer meant here. 13. Line 1: "an algorithm" --> "models" Done. 14. Sentence 1: Begin the paragraph with: AI-driven systems could have real potential to improve stability, predictability ... Adust the paragraph appropriately. We rephrased the sentence. See lines 89-95. 15. While the benefits of such systems are described - it must be mentioned that much more work is required before automated systems replace human decision making. Done. See lines 96-99. 16. INSTITUTIONAL BACKGROUD - This section must be reduced - while it may be usefull background - much of it is not necessary to understand the rest of the paper. It should rather be reduced from 5 pages to 2 or so pages. We reduced this section from 2545 to 1700 words (a 33% decrease from the original size). We could not reach the size by you without losing too much information. We believe it is essential to put things in context by explaining where our models should be deployed. While there are many scientific papers about LSTMs and Transformers in English, there is not enough material about the Brazilian legal system. As legal researchers, we feel obliged to provide this information to our readers, as we do not expect our audience to understand how Brazilian courts work. 17. NLP MODELS - The paper should include a background on the NLP models used in the study. For e.g. LSTM and Transformer models should be described. Done. See lines 234-342. 18. METHODOLOGY - Paragraph Two: clarify if any pre-processing was done - either no preprocessing, or if any then what did this entail. Done. See lines 363-366. 19. Paragraph Three: did each participant predict each of the 690 appeals? Done. See lines 369-390 and 609. 20. In what language are the full text written in? Done. See lines 366-367. 21. Page 13 - first two paragraphs - would better fit in the introduction that discusses the motivation for the study. Done. See lines 64-77. 22. Page 14: While the labelling process is sensible, it is not fool proof. Where there any processes used to verify the labels? For e.g. sampling some of the decisions? Done. See lines 448-449. 23. Page 16: First paragraph in "Experimental setup" ".. to all parties in production. " --> "... to all parties." Done. 24. Experimental Setup: The training set up of the different models should give the hyper parameters used and explain how these were tuned. Done. See lines 510-514 and 514-589. 25. Some more details of the model set up would be useful if they differed from the standard set up discussed in the literature. Done. See lines 510-589, especially (new) tables 2, 3, and 4. 26. Legal expert's analysis"... 17 highly skilled ..." This number contradicts the numbers given earlier in the paper - where 22 is reported. Please fix. Done. 27. Page 19: While we understand the difficulty to obtain human expert decisions,it would have good to evaluate inter-human agreement i.e. if participants were to have evaluated the same cases. We understand and agree with your point. However, we knew how busy our subjects were with their everyday court activities. Therefore, asking them to label the same data point more than once would be risky, resulting in less labeled data on our human expert’s dataset, weakening the comparison between humans and ML models (our primary goal). 28. RESULTS - The hyperparameters must be clearly specified in the methods. Done. Please see tables 2, 3 and 4 (lines 531, 563, and 588) 29. DISCUSSION - Paragraph Two: Line 2: "example The" --> "examople. The" Done. 33. Paragraph Four: Sentence Two is not substantiated: How will the tool help to detect inconsistencies - as the tool does not, as described, cluster cases on similarity. So this entire paragraph is problematic and should be left out or substantiated carefully. We removed this part. Lastly, we want to thank you for the thorough review of our work. C) REVIEWER 2 1.The paper provides an interesting contribution to the literature. I believe that the paper could improve the explanation of the models that are used and provide the codes so other researchers can follow these steps. It could open an important reseach area applied to judicial sentences and increase the potential for citations of the paper. Thanks for your input. We provided additional information about the architectures and hyperparameters used. See lines 510-589, especially (new) tables 2, 3, and 4. All code is available at https://github.com/eliasjacob/paper_brcad5 Attachments Attachment Submitted filename: Response to Reviewers.docx https://doi.org/10.1371/journal.pone.0272287.r002
18 Jul 2022 Decision Letter - Donrich Thaldar, Editor Using deep learning to predict outcomes of legal appeals better than human experts: A study with data from Brazilian federal courts PONE-D-22-07895R1 Dear Dr. Jacob de Menezes-Neto, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, Donrich Thaldar Academic Editor PLOS ONE Additional Editor Comments (optional): Reviewers' comments: https://doi.org/10.1371/journal.pone.0272287.r003
Formally Accepted
20 Jul 2022 Acceptance Letter - Donrich Thaldar, Editor PONE-D-22-07895R1 Using deep learning to predict outcomes of legal appeals better than human experts: A study with data from Brazilian federal courts Dear Dr. Jacob de Menezes-Neto: I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department. If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org. If we can help with anything else, please email us at plosone@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Professor Donrich Thaldar Academic Editor PLOS ONE https://doi.org/10.1371/journal.pone.0272287.r004

Open letter on the publication of peer review reports

PLOS recognizes the benefits of transparency in the peer review process. Therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. Reviewers remain anonymous, unless they choose to reveal their names.

We encourage other journals to join us in this initiative. We hope that our action inspires the community, including researchers, research funders, and research institutions, to recognize the benefits of published peer review reports for all parts of the research system.

Learn more at ASAPbio .