Legal large language models (LLMs) deployed in high-stakes judicial settings must exhibit robustness against non-substantive linguistic variations while preserving acute sensitivity to legally determinative facts and norms. This study investigates this robustness–reliability trade-off within the context of Chinese legal tasks. We curate a dataset of 5,000 Chinese judicial question–answer pairs and generate semantic-preserving adversarial rewrites, retaining only those validated by an embedding-based semantic consistency filter. Holding the total training budget and fine-tuning protocol constant, we fine-tune model variants that differ exclusively (...) in their injection ratio of these verified rewrites, establishing seven distinct injection groups (G0–G6). We evaluate model reliability utilizing a composite protocol that incorporates objective accuracy on exam-style questions, expert evaluations of open-ended responses, and embedding-based semantic similarity. For trademark infringement reasoning tasks, we additionally assess verdict accuracy and rationale quality. Across varying model capacities (4B, 20B, and 32B backbones, including Qwen3-4B, GPT-OSS-20B, and Qwen3-VL-32B-Instruct) and both evaluated tasks, our findings reveal an inverted‑U relationship: moderate robustness data injection enhances reliability, whereas excessive injection degrades overall performance and induces characteristic failure modes, such as the attenuation of legally salient distinctions, the generation of boilerplate rationales, and overly cautious abstention. These findings substantiate “moderate robustness injection” as a practical heuristic and underscore a broader principle of differential sensitivity—achieving insensitivity to superficial variations without blunting the model’s sensitivity to legally decisive elements. (shrink)

In recent years, supply chain finance is exploited to solve the financing difficulties of small- and medium-sized enterprises. SME credit risk assessment is a critical part in the SCF system. The diffusion of SME credit risk may cause serious consequences, leading the whole supply chain finance system unstable and insecure. Compared with traditional credit risk assessment models, the supply chain relationship, credit condition of SME, and core enterprises should all be considered to rate SME credit risk in SCF. Traditional methods (...) mix all indicators from different index systems. They cannot give a quantitative result on how these index systems work. Furthermore, traditional credit risk assessment models are heavily dependent on the number of annotated SME data. However, it is implausible to accumulate enough credit risky SMEs in advance. In this paper, we propose an adaptive heterogenous multiview graph learning method to tackle the small sample size problem for SMEs’ credit risk forecasting. Three graphs are constructed by using indicators from supply chain operation, SME financial indicator, and nonfinancial indicator individually. All the graphs are integrated in an adaptive manner, providing a quantitative explanation on how the three parts cooperate. The experimental analysis shows that the proposed method has good performance for determining whether SME is risky or nonrisky in SCF. From the perspective of SCF, SME financing ability is still the main factor to determine the credit risk of SME. (shrink)

Reseña de la tesis doctoral del mismo título de Yan Shuhua.

Reputation detection in traditional distributed systems relies heavily on the factor of interaction reputation, which can be derived from direct interactions between agents through bidirectional relationships. However, in the current information diffusion in social network systems, the characteristic of the unidirectional relationship between agents and the decay property of diffusion will result in lacking direct interactions; therefore, interaction reputations will be difficult to be obtained by agents in a distributed manner. To solve this problem, a novel distributed reputation detection model (...) following the pattern “from path to individual” is proposed, which can provide a new reputation factor as an alternative to interaction reputation in such environments. The main idea is that the positive observation of an information diffusion process increases the belief of the corresponding diffusion path, which further increases the reputation of each involved agent. Thus, the reputation of a target agent can be assessed by the superimposition of reputations of multiple paths on which this agent has participated in past information diffusion processes. Furthermore, being aware of agent’s limited capacity for reputation detection in SNS, we then propose the enhanced FPTI model, which simplifies the detection source to reduce detection costs and achieve the approximate performance as FPTI. Theoretical analyses and experimental evaluations validate the efficiency and effectiveness of our models and also show several properties of the models, for example, the robustness for dynamic environments. (shrink)

In 2011, Jiangsu Province set up the guidance fund for sports industry. In 2016, Jiangsu Provincial Finance Department jointly issued the document with Jiangsu Provincial Sports Bureau, which was changed to “Jiangsu Special Fund for Sports Industry Development.” By 2019, Jiangsu Province has funded 1,009 sports industry development projects with a total investment of 796.6 million yuan. The effect of the investment has attracted extensive attention from the government, society, and enterprises. Entrusted by Jiangsu Sports Industry Guidance Center, this research (...) team spent one year on the performance research of 1,009 projects supported by special funds for the development of sports industry, from the construction of evaluation index system to the design, issuance, collection, and processing of questionnaires. Using the methods of literature, questionnaire, expert interview, and mathematical statistics, this paper takes the performance management of the special funds for the development of sports industry in Jiangsu Province as the research object, based on the investigation and analysis of the current situation of the special funds for the development of sports industry in Jiangsu Province. It analyzes the special funds from four aspects: the rate of funds in place, the ratio of funds pulled, the completion rate of project implementation, and the rate of fund prerecovery. Then, this paper expounds the current situation, effectiveness, and problems of the performance management of the special funds for the development of sports industry and puts forward some countermeasures and suggestions for improvement. (shrink)

Resumen: En China los templos de varias creencias o religiones conforman una parte importante del espacio sagrado. En este trabajo pretendemos realizar una introducción a los templos de las creencias prehistóricas, ligadas a la naturaleza y los ancestros, así como a los de las tres principales religiones en dicho país, que son el taoísmo, el budismo y el confucianismo. Por otra parte, se exponen de manera sucinta la ubicación, el diseño, la historia, los distintos modelos y las deidades consagradas de (...) los templos. (shrink)

Cognitive neuroscience research has found that individuals from different cultures have different neural responses and emotional perceptions. Differences in executives’ perception of external pay gaps in different cultures can affect their work attitudes and behavior. In this study, we explore the direct relationship between executive compensation fairness and executive innovation motivation. We also investigate the moderating effects of Confucian culture and western culture between executive compensation fairness and executive innovation motivation. Data were collected from the Chinese listed firms from 2011 (...) to 2019 and test the relationship using the fixed-effect panel regression models. The results demonstrate that executive compensation fairness positively influences the executive innovation motivation. This effect is more pronounced in Confucian culture regions, while Western culture weakens this effect. The findings of this study confirmed that executive compensation fairness, provide a cross-cultural comparison for compensation research, validate the findings of cultural neuroscience, and provide useful insights into the research of common prosperity. To improve the corporate compensation structure, it is necessary to consider the relative pay equity with firms in the same region and the influence of corporate culture. (shrink)