Abstract
Legal Judgment Prediction (LJP) is an essential component of legal assistant systems, which aims to automatically predict judgment results from a given criminal fact description. As a vital subtask of LJP, researchers have paid little attention to the numerical LJP, i.e., the prediction of imprisonment and penalty. Existing methods ignore numerical information in the criminal facts, making their performances far from satisfactory. For instance, the amount of theft varies, as do the prison terms and penalties. The major challenge is how the model can obtain the ability of numerical comparison and magnitude perception, e.g., 400 < 500 < 800, 500 is closer to 400 than to 800. To this end, we propose a judicial knowledge-enhanced magnitude-aware reasoning architecture, called NumLJP, for the numerical LJP task. Specifically, we first implement a contrastive learning-based judicial knowledge selector to distinguish confusing criminal cases efficiently. Unlike previous approaches that employ the law article as external knowledge, judicial knowledge is a quantitative guideline in real scenarios. It contains many numerals (called anchors) that can construct a reference frame. Then we design a masked numeral prediction task to help the model remember these anchors to acquire legal numerical commonsense from the selected judicial knowledge. We construct a scale-based numerical graph using the anchors and numerals in facts to perform magnitude-aware numerical reasoning. Finally, the representations of fact description, judicial knowledge, and numerals are fused to make decisions. We conduct extensive experiments on three real-world datasets and select several competitive baselines. The results demonstrate that the macro-F1 of NumLJP improves by at least 9.53% and 11.57% on the prediction of penalty and imprisonment, respectively.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
These numerals may change as the legal system is reformed, but they are fixed over a considerable period. Therefore, we assume that these numerals are fixed.
Legal numerical commonsense indicates the judge’s knowledge of the numerical features in the fact description, such as the amount of property stolen, the number of drugs sold, etc. Each of these numerals has its own range and probability distribution.
Here the numeral vocabulary refers to all numerical anchors that appear in a same judicial knowledge.
Existing Chinese LJP datasets are usually divided in this manner.
PLMs utilize called WordPiece tokenizer to split words either into the full forms or into word pieces Devlin et al. (2019).
the anchors of Theft are 1,000, 3,000, 30,000, 100,000, 300,000, 500,000.
Among all hyperparameters, the learning rate lr, gradient clipping clipping, the weight of contrastive learning loss \(\lambda \), and the temperature \(\tau \) are set empirically following previous works, which are not repeated in this paper. \(N^t\) is the multiplier assigned for interval division, and we detail its setting principle in Section 4.3.1.
The comparison chain is ordered numerals in a numerical graph.
References
Amini A, Gabriel S, Lin S, Koncel-Kedziorski R, Choi Y, Hajishirzi H (2019) Mathqa: Towards interpretable math word problem solving with operation-based formalisms. In: NAACL, pp. 2357–2367
Bakalov A, Fuxman A, Talukdar PP, Chakrabarti S (2011) Scad: Collective discovery of attribute values. In: WWW, pp. 447–456
Baly R, Karadzhov G, Saleh A, Glass JR, Nakov P (2019) Multi-task ordinal regression for jointly predicting the trustworthiness and the leading political ideology of news media. In: NAACL-HLT, pp. 2109–2116
Banerjee S, Chakrabarti S, Ramakrishnan G (2009) Learning to rank for quantity consensus queries. In: SIGIR, pp. 243–250
Bi S, Huang Y, Cheng X, Wang M, Qi G (2019) Building chinese legal hybrid knowledge network. KSEM 11775:628–639
Bi S, Cheng X, Chen J, Qi G, Wang M, Zhou Y, Wang L (2019) Dispute generation in law documents via joint context and topic attention. In: JIST, pp. 116–129
Brown T, Mann B, Ryder N, Subbiah M, Kaplan JD, Dhariwal P, Neelakantan A, Shyam P, Sastry G, Askell A et al (2020) Language models are few-shot learners. Neural Inf Process Syst 33:1877–1901
Cao W, Mirjalili V, Raschka S (2020) Rank consistent ordinal regression for neural networks with application to age estimation. Pattern Recognit Lett 140:325–331
Chalkidis I, Androutsopoulos I, Aletras N (2019) Neural legal judgment prediction in English. In: ACL, pp. 4317–4323
Chen H, Cai D, Dai W, Dai Z, Ding Y (2019) Charge-based prison term prediction with deep gating network. In: EMNLP, pp. 6361–6366
Chen K, Xu W, Cheng X, Xiaochuan Z, Zhang Y, Song L, Wang T, Qi Y, Chu W (2020) Question directed graph attention network for numerical reasoning over text. In: EMNLP, pp. 6759–6768
Cheng X, Bi S, Qi G, Wang Y (2020) Knowledge-aware method for confusing charge prediction. NLPCC 12430:667–679
Devlin J, Chang M, Lee K, Toutanova K (2019) BERT: pre-training of deep bidirectional transformers for language understanding. In: NAACL, pp. 4171–4186
Diaz R, Marathe A (2019) Soft labels for ordinal regression. In: CVPR, pp. 4738–4747
Dong Q, Niu S (2021) Legal judgment prediction via relational learning. In: SIGIR, pp. 983–992
Dua D, Wang Y, Dasigi P, Stanovsky G, Singh S, Gardner M (2019) DROP: A reading comprehension benchmark requiring discrete reasoning over paragraphs. In: NAACL, pp. 2368–2378
Fleiss JL (1971) Measuring nominal scale agreement among many raters. Psychol Bull 76(5):378
Ge J, Huang Y, Shen X, Li C, Hu W (2021) Learning fine-grained fact-article correspondence in legal cases. TASLP 29:3694–3706
George TE, Epstein L (1992) On the nature of supreme court decision making. APSR 86(2):323–337
Geva M, Gupta A, Berant J (2020) Injecting numerical reasoning skills into language models. In: ACL, pp. 946–958
Gunel B, Du J, Conneau A, Stoyanov V (2021) Supervised contrastive learning for pre-trained language model fine-tuning. In: ICLR
Guo Z, Zhang Y, Teng Z, Lu W (2019) Densely connected graph convolutional networks for graph-to-sequence learning. TACL 7:297–312
Gutmann M, Hyvärinen A (2010) Noise-contrastive estimation: a new estimation principle for unnormalized statistical models. AISTATS 9:297–304
Hamilton WL, Ying Z, Leskovec J (2017) Inductive representation learning on large graphs. In: NeurIPS, pp. 1024–1034
Hu Z, Li X, Tu C, Liu Z, Sun M (2018) Few-shot charge prediction with discriminative legal attributes. In: COLING, pp. 487–498
Huang D, Shi S, Lin C, Yin J, Ma W (2016) How well do computers solve math word problems? large-scale dataset construction and evaluation. In: ACL
Huber PJ (1992) Robust estimation of a location parameter. In: Breakthroughs in Statistics, pp. 492–518
Hénaff OJ (2020) Data-efficient image recognition with contrastive predictive coding. ICML 119:4182–4192
Jaiswal A, Babu AR, Zadeh MZ, Banerjee D, Makedon F (2021) A survey on contrastive self-supervised learning. Technologies 9(1):2
Jiang C, Nian Z, Guo K, Chu S, Zhao Y, Shen L, Tu K (2019) Learning numeral embeddings. arXiv preprint arXiv:2001.00003
Khosla P, Teterwak P, Wang C, Sarna A, Tian Y, Isola P, Maschinot A, Liu C, Krishnan D (2020) Supervised contrastive learning. Neural Inf Process Syst, 33
Kingma DP, Ba J (2015) Adam: A method for stochastic optimization. In: ICLR
Kort F (1957) Predicting supreme court decisions mathematically: a quantitative analysis of the “right to counsel’’ cases. APSR 51(1):1–12
Lewis M, Liu Y, Goyal N, Ghazvininejad M, Mohamed A, Levy O, Stoyanov V, Zettlemoyer L (2020) BART: denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. In: ACL, pp. 7871–7880
Li S, Zhang H, Ye L, Su S, Guo X, Yu H, Fang B (2020) Prison term prediction on criminal case description with deep learning. Comput Mater Contin 62(3):1217–1231
Lin BY, Lee S, Khanna R, Ren X (2020) Birds have four legs?! numersense: Probing numerical commonsense knowledge of pre-trained language models. In: EMNLP, pp. 6862–6868
Liu YH, Chen YL, Ho WL (2015) Predicting associated statutes for legal problems. IPM 51(1):194–211
Liu C-L, Chang C-T, Ho J-H (2004) Case instance generation and refinement for case-based criminal summary judgments in chinese. JISE, 783–800
Liu CL, Liao TM (2005) Classifying criminal charges in chinese for web-based legal services. In: APCCMI
Liu Y, Ott M, Goyal N, Du J, Joshi M, Chen D, Levy O, Lewis M, Zettlemoyer L, Stoyanov V (2019) Roberta: A robustly optimized BERT pretraining approach. CoRR abs/1907.11692
Luo B, Feng Y, Xu J, Zhang X, Zhao D (2017) Learning to predict charges for criminal cases with legal basis. In: EMNLP, pp. 2727–2736
Nie Y, Williams A, Dinan E, Bansal M, Weston J, Kiela D (2020) Adversarial NLI: A new benchmark for natural language understanding. In: ACL, pp. 4885–4901
Niu Z, Zhou M, Wang L, Gao X, Hua G (2016) Ordinal regression with multiple output CNN for age estimation. In: CVPR, pp. 4920–4928
Parikh N, Boyd SP (2014) Proximal algorithms. Found. Trends Optim. 1(3):127–239
Patel A, Bhattamishra S, Goyal N (2021) Are NLP models really able to solve simple math word problems? In: NAACL, pp. 2080–2094
Qin J, Lin L, Liang X, Zhang R, Lin L (2020) Semantically-aligned universal tree-structured solver for math word problems. In: EMNLP, pp. 3780–3789
Ran Q, Lin Y, Li P, Zhou J, Liu Z (2019) Numnet: Machine reading comprehension with numerical reasoning. In: EMNLP, pp. 2474–2484
Ribeiro MT, Wu T, Guestrin C, Singh S (2020) Beyond accuracy: Behavioral testing of NLP models with checklist. In: ACL, pp. 4902–4912
Robinson J.D, Chuang C, Sra S, Jegelka S (2021) Contrastive learning with hard negative samples. In: ICLR
Saha A, Joty SR, Hoi SCH (2021) Weakly supervised neuro-symbolic module networks for numerical reasoning. CoRR abs/2101.11802
Sanh V, Debut L, Chaumond J, Wolf T (2019) Distilbert, a distilled version of BERT: smaller, faster, cheaper and lighter. CoRR abs/1910.01108
Segal JA (1984) Predicting supreme court cases probabilistically: The search and seizure cases, 1962-1981. APSA 78
Sermanet P, Lynch C, Chebotar Y, Hsu J, Jang E, Schaal S, Levine S (2018) Time-contrastive networks: Self-supervised learning from video. In: ICRA, pp. 1134–1141
Shi X, Cao W, Raschka S (2021) Deep neural networks for rank-consistent ordinal regression based on conditional probabilities. CoRR abs/2111.08851
Shorten C, Khoshgoftaar TM, Furht B (2021) Text data augmentation for deep learning. J Big Data 8(1):101
Spithourakis GP, Riedel S (2018) Numeracy for language models: Evaluating and improving their ability to predict numbers. In: ACL, pp. 2104–2115
Thawani A, Pujara J, Ilievski F, Szekely PA (2021) Representing numbers in NLP: a survey and a vision. In: NAACL, pp. 644–656
Tian Y, Krishnan D, Isola P (2020) Contrastive multiview coding. In: ECCV, vol. 12356, pp. 776–794. Springer
Van der Maaten L, Hinton G (2008) Visualizing data using t-sne. JMLR 9(11)
van den Oord A, Li Y, Vinyals O (2018) Representation learning with contrastive predictive coding. CoRR abs/1807.03748
Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser L, Polosukhin I (2017) Attention is all you need. Neural Inf Process Syst, pp. 5998–6008
Wu Z, Xiong Y, Yu SX, Lin D (2018) Unsupervised feature learning via non-parametric instance discrimination. In: CVPR, pp. 3733–3742
Xiao C, Zhong H, Guo Z, Tu C, Liu Z, Sun M, Feng Y, Han X, Hu Z, Wang H, Xu J (2018) CAIL2018: A large-scale legal dataset for judgment prediction. CoRR abs/1807.02478
Xu N, Wang P, Chen L, Pan L, Wang X, Zhao J (2020) Distinguish confusing law articles for legal judgment prediction. In: ACL, pp. 3086–3095
Yang W, Jia W, Zhou X, Luo Y (2019) Legal judgment prediction via multi-perspective bi-feedback network. In: IJCAI, pp. 4085–4091
Yoran O, Talmor A, Berant J (2022) Turning tables: Generating examples from semi-structured tables for endowing language models with reasoning skills. In: ACL, pp. 6016–6031
Yue L, Liu Q, Jin B, Wu H, Zhang K, An Y, Cheng M, Yin B, Wu D (2021) Neurjudge: A circumstance-aware neural framework for legal judgment prediction. In: SIGIR, pp. 973–982
Zhong H, Guo Z, Tu C, Xiao C, Liu Z, Sun M (2018) Legal judgment prediction via topological learning. In: EMNLP, pp. 3540–3549
Zhong H, Xiao C, Tu C, Zhang T, Liu Z, Sun M (2020) How does NLP benefit legal system: A summary of legal artificial intelligence. In: ACL, pp. 5218–5230
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Bi, S., Zhou, Z., Pan, L. et al. Judicial knowledge-enhanced magnitude-aware reasoning for numerical legal judgment prediction. Artif Intell Law 31, 773–806 (2023). https://doi.org/10.1007/s10506-022-09337-4
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10506-022-09337-4