Logic-Guided Multistage Inference for Explainable Multidefendant Judgment Prediction

TL;DR

This work tackles the challenge of predicting guilt and sentencing in multidefendant cases, where defendant roles (principal vs accomplice) critically influence outcomes. It introduces MMSI, a two-stage Transformer framework that performs guilt inference from factual descriptions and then sentencing regression using a pruned court view, with an oriented masking module and label broadcasting to fuse role signals. Key contributions include defendant-oriented masking, comparative data construction for robust role differentiation, a label broadcasting mechanism, and the IMLJP dataset with extensive annotations; MMSI achieves state-of-the-art results and competitive performance against SOTA LLMs. The approach offers interpretable, logic-guided judicial assistance and is extensible to other crimes and jurisdictions.

Abstract

Crime disrupts societal stability, making law essential for balance. In multidefendant cases, assigning responsibility is complex and challenges fairness, requiring precise role differentiation. However, judicial phrasing often obscures the roles of the defendants, hindering effective AI-driven analyses. To address this issue, we incorporate sentencing logic into a pretrained Transformer encoder framework to enhance the intelligent assistance in multidefendant cases while ensuring legal interpretability. Within this framework an oriented masking mechanism clarifies roles and a comparative data construction strategy improves the model's sensitivity to culpability distinctions between principals and accomplices. Predicted guilt labels are further incorporated into a regression model through broadcasting, consolidating crime descriptions and court views. Our proposed masked multistage inference (MMSI) framework, evaluated on the custom IMLJP dataset for intentional injury cases, achieves significant accuracy improvements, outperforming baselines in role-based culpability differentiation. This work offers a robust solution for enhancing intelligent judicial systems, with publicly code available.

Figures (12)

• Figure 1: Framework of the proposed MMSI for multidefendant cases. This approach involves two primary tasks: Task 1 (shown in red), which performs defendant guilt--responsibility reasoning on the basis of fact descriptions (FDs), and Task 2 (shown in blue), which predicts sentencing terms via both FDs and pruned court views (CV$_d$, i.e., court views with guilt-role-related sentences removed). Both the FD and CV$_d$ inputs undergo preprocessing with an oriented masking technique, and the training data are enhanced through contrastive data construction, which is applied to both tasks to achieve improved prediction accuracy. In Task 2, the guilt responsibility labels derived from Task 1 are broadcast within the model and embedded into text representations, facilitating the integration of multisource data into MMSI for prison prediction.
• Figure 2: Structures of judicial documents and sentencing logic. The left panel illustrates the temporal sequence consisting of the crime, trial, and sentencing stages. The middle panel presents the composition of a judicial document. The right panel outlines the sentencing logic process for determining prison terms under a specific criminal charge.
• Figure 3: Example of the oriented masking strategy. Based on the target defendant contained in the input, the left panel illustrates the masking process applied to the FD, whereas the right panel applies masking to the CV for the same defendant.
• Figure 4: The architecture of the proposed MMSI framework and its correspondence to real-world judicial procedures. The left panel illustrates the end-to-end structure of MMSI, which jointly performs guilt inference and prison prediction based on three textual inputs: Name, FD, and CV$_{d}$. Through oriented masking, the model isolates the target defendant in multidefendant cases and infers guilt from FD using a classification branch. The inferred guilt label is then integrated into a regression branch that combines CV$_{d}$ and Name to predict the imprisonment term. The Embedding modules refer to the word-level embeddings produced by BERT tokenizer and its input embedding layer. The right panel depicts the real-world judicial reasoning process that MMSI emulates, including guilt inference from FD and sentencing based on CV$_{d}$. The oriented masking mechanism serves as a computational proxy for the judge’s focus on the target defendant.
• Figure 5: Defendant distribution in the IMLJP dataset. a) represents the number of defendants per case and their distribution (in blue), b) illustrates the proportion of defendants relative to all individuals involved in each case (in orange), and c) depicts the proportion of defendants identified as principal offenders (in green).