A concurrent prediction of criminal law charge and sentence using twin convolutional neural networks

An intelligent law article prediction scheme, which solves the law articles imbalance problem and the missing value problem of the judgement, is proposed in this paper. This paper applies the law article description as the label attribute. Based on the property of the vector space, the missing value problem can be got over by learning a representative embedding vector through the vector similarity weighted mechanism. For the imbalance problem, we use a weight sharing classification layer which classifies the label according to the relevance between the fact vector and the law article vector of the vector space. We also use the transfer learning to train the model by the high-frequency law articles first, then share the weight as the prior knowledge to the low-frequency one to improve the classification performance. The proposed approach outperforms the performance on few-shot law article prediction.