Legal Judgment Prediction aims to automatically predict judgment outcomes based on descriptions of legal cases and established law articles, and has received increasing attention. In the preliminary work, several problems still have not been adequately solved. One is how to utilize limited but valuable label information. Existing methods mostly ignore the gap between the description of established articles and cases, but directly integrate them. Second, most studies ignore the mutual constraint among the subtasks, such as logically or semantically, each charge is only related to some specific articles. To address these issues, we first construct a crime similarity graph and then perform a distillation operation to collect discriminate keywords for each charge. Furthermore, we fuse these discriminative keywords instead of established article descriptions into case embedding with a cross-attention mechanism to obtain deep semantic representations of cases incorporating label information. Finally, under a constraint among subtasks, we optimize the one-hot representation of ground-truth labels to guarantee consistent results across the subtasks based on the label-enhancement algorithm. To verify the effectiveness and robustness of our framework, we conduct extensive experiments on two public datasets. The experimental results show that the proposed method outperforms the state-of-art models by 3.89%/7.92% and 1.23%/2.50% in the average MF1-score of the subtasks on CAIL-Small/Big, respectively.

As one of the most critical tasks in legal artificial intelligence, legal judgment prediction (LJP) has garnered growing attention, especially in the civil law system. However, current methods often overlook the challenge of imbalanced label distributions, treating each label with equal importance, which can lead the model to be biased toward labels with high frequency. In this paper, we propose a label-enhanced prototypical network (LPN) suitable for LJP, that adopts a strategy of uniform encoding and separate decoding. Specifically, LPN adopts a multi-scale convolutional neural network to uniformly encode case factual description to capture long-distance features of the document. At the decoding end, a prototypical network incorporating label semantic features is used to guide the learning of prototype representations of high-frequency and low-frequency labels, respectively. At the same time, we also propose a prototype-prototype loss to optimize the prototypical representation. We conduct extensive experiments on two real datasets and show that our proposed method effectively improves the performance of LJP, with an average F1 of 1.23% and 1.13% higher than the state-of-the-art model on two subtasks, respectively.

Charge prediction aims to predict charges from the case descriptions and plays a significant role in legal assistance systems. When we use deep learning methods, prediction on high-frequency charges has achieved promising results, but prediction on few-shot charges is still a challenge. To address this issue, a few-shot charge prediction method with data augmentation and neural network is proposed, named BBAE (BERT-BiGRU-Attention based on easy data augmentation techniques), which can be divided into three layers: data augmentation layer, encoder layer, and output layer. Specifically, the data augmentation layer takes the case description as input and uses EDA (easy data augmentation techniques) to generate synthetic samples biased to few-shot charges based on charge categories. The encoder layer employs the BERT-BiGRU-Attention model to fully extract text features, while the output layer predicts the charge on the basis of text features. Experiments on three public datasets of Chinese criminal cases demonstrate that our method achieves more effective improvements over other baselines. BBAE outperforms state-of-the-art methods by 4.6% and 9.3% under Macro F1 in low-frequency and medium-frequency charges, which indicates that our method is effective in few-shot charge prediction.

Emotion Recognition in Conversation (ERC) aims to recognize the emotion for each utterance in a conversation automatically. Due to the difficulty of collecting and labeling, this task lacks the dataset corpora available on a large scale. This increases the difficulty of finishing the supervised training required by large-scale neural networks. Introducing the large-scale generative conversational dataset can assist with modeling dialogue. However, the spatial distribution of feature vectors in the source and target domains is inconsistent after introducing the external dataset. To alleviate the problem, we propose a Domain Adversarial Network for Cross-Domain Emotion Recognition in Conversation (DAN-CDERC) model, consisting of domain adversarial and emotion recognition models. The domain adversarial model consists of the encoders, a generator and a domain discriminator. First, the encoders and generator learn contextual features from a large-scale source dataset. The discriminator performs domain adaptation by discriminating the domain to make the feature space of the source and target domain consistent, so as to obtain domain invariant features. Then DAN-CDERC transfers the learned domain invariant dialogue context knowledge from the domain adversarial model to the emotion recognition model to assist in modeling the dialogue context. Due to the use of a domain adversarial network, DAN-CDERC obtains dialogue-level contextual information that is domain invariant, thereby reducing the negative impact of inconsistency in domain space. Empirical studies illustrate that the proposed model outperforms the baseline models on three benchmark emotion recognition datasets.

Legal Article Prediction (LAP) aims to automatically identify the relevant articles according to the content of the prosecution document of each case and the established law articles, which is of great value in legal assistance systems and in improving the work efficiency of legal practitioners. In this paper, we formulate LAP as a multi-label learning problem and present a Multiple Attention Label-Enhanced Model (MALE), which first utilizes two encoders to obtain the base representation of fact description and established law articles, second employs a labelenhanced method fully fusing the information of the two from different perspectives to get the co-dependent representation of the fact description and each law article finally. Experimental results on the CAIL2018 public dataset demonstrate that MALE achieves significant performance improvements over existing neural models in predicting relevant law articles.

Legal judgment prediction (LJP) is used to predict judgment results based on the description of individual legal cases. In order to be more suitable for actual application scenarios in which the case has cited multiple articles and has multiple charges, we formulate legal judgment prediction as a multiple label learning problem and present a deep learning model that can effectively encode the content of each legal case via a multi-residual convolution neural network and the semantics of law articles via an article encoder. An article-wise attention mechanism is proposed to fuse the two types of encoded information. Experimental results derived on the CAIL2018 datasets show that our model provides a significant performance improvement over the existing neural models in predicting relevant law articles and charges.

Charge prediction aims to predict the corresponding charges for a specific case. In civil law system, human judges will match the facts with relevant laws, and the final judgments are usually made in accordance with relevant law articles. Existing works either ignore this feature or simply model the relationship using multi-task learning, but neither make full use of relevant articles to assist the charge prediction task. To address this issue, we propose an attentional neural network, LegalAtt, which uses relevant articles to improve the performance and interpretability of charge prediction task. More specifically, our model works in a bidirectional approach: First, it uses the fact description to extract relevant articles; In return, the selected relevant articles assist to locate key information from the fact description, which helps improve the performance of charge prediction. Experimental results show that our model achieves the best performance on the real-world dataset compared with other state-of-the-art baselines. Our code is available at https://github.com/nlp208/legal_attention.

The key problem of supervising word sense disambiguation is the lack of a large-scale and high-quality corpus of word sense tagging. Based on the Contemporary Chinese Semantic Dictionary, the Modern Chinese Dictionary (5th Edition) and the Chinese Lexical Semantic Knowledge Base, this paper analyzes the adjectives, nouns and verbs with polysemic in the dictionaries and fuses them together to construct the Zhengzhou University Contemporary Chinese Semantic Dictionary. People's Daily corpus is selected for annotation, and the word sense tagging corpus with 1.87 million words is constructed. It is expected to provide better data support for natural language processing tasks such as semantic automatic analysis and word sense disambiguation. This paper presents a detailed and rigorous specification of word sense tagging in the process of annotation. In addition, in the new domain corpus, the automatic annotation method achieved excellent performance, which can be used for subsequent reference.