数字化社会治理的复杂性机理与智能方法专题
信息-物理-社会系统与数智时代社会治理
王重阳, 黄棣芳, 朱欣怡, 郝俊, 蔡世民, 杨晓光, 刘志新, 王芳, 李建平, 吕琳媛
中国科学: 信息科学, 2026, 56(9): 2203-2227
关键词 数智化社会; 信息-物理-社会系统; 社会治理; 复杂性科学; 复杂系统; digital intelligent society; cyber-physical-social system; social governance; complexity science; complex system
Chong-Yang WANG, Difang HUANG, Xinyi ZHU, et al. Cyber-physical-social system and social governance in the digital-intelligent era. Sci Sin Inform, 2026, 56(9): 2203-2227, doi: 10.1360/SSI-2026-0207
数字化社会治理的复杂性机理与智能方法专题
基于深度强化学习的数据产品动态定价: 一种多头注意力机制的近端策略优化方法
李建平, 尚书凡, 郝俊, 林祺, 李金
中国科学: 信息科学, 2026, 56(9): 2228-2245
关键词 数据定价; 场内交易; 深度强化学习; PPO; 多头注意力机制; data pricing; exchange trading; deep reinforcement learning; multi-head attention mechanism
Jianping LI, Shufan SHANG, Jun HAO, et al. Dynamic pricing for data products based on deep reinforcement learning: a multi-head attention proximal policy optimization approach. Sci Sin Inform, 2026, 56(9): 2228-2245, doi: 10.1360/SSI-2026-0060
数字化社会治理的复杂性机理与智能方法专题
多被告缓刑适用预测: 基于语义-结构融合与类案知识增强的两步自适应分类方法
戴瑞芬, 杨玲燕, 王芳
中国科学: 信息科学, 2026, 56(9): 2246-2261
关键词 多被告缓刑适用预测; 两步自适应分类; 加权最小二乘; 非独立同分布数据; 语义-结构融合; multi-defendant probation applicability prediction; two-step adaptive classification; weighted least squares; non-i.i.d. data; semantic-structural fusion
Ruifen DAI, Lingyan YANG, Fang WANG. Multi-defendant probation applicability prediction: a two-step adaptive classification method based on semantic-structural fusion and similar-case knowledge augmentation. Sci Sin Inform, 2026, 56(9): 2246-2261, doi: 10.1360/SSI-2026-0174
数字化社会治理的复杂性机理与智能方法专题
基于CPSS框架的突发事件多主体演化博弈与协同治理
郭金忠, 刘雅婕, 刘晓玲, 康明与, 吕琳媛
中国科学: 信息科学, 2026, 56(9): 2262-2282
关键词 CPSS; 突发事件; 演化博弈; 信息治理; 协同治理; emergencies; evolutionary game; information governance; collaborative governance
Jinzhong GUO, Yajie LIU, Xiaoling LIU, et al. Multi-agent evolutionary game and synergistic governance of emergency events: a CPSS-based approach. Sci Sin Inform, 2026, 56(9): 2262-2282, doi: 10.1360/SSI-2026-0230
数字化社会治理的复杂性机理与智能方法专题
饱和观测下逐分量自适应预测理论及其分阶段耦合量刑应用
刘宇静, 金依菲, 刘志新, 郭雷
中国科学: 信息科学, 2026, 56(9): 2283-2309
关键词 量刑预测; 缓刑预测; 时变饱和模型; 逐分量自适应似然梯度算法; 分阶段耦合量刑模型; 盗窃罪; sentencing prediction; probation prediction; time-varying saturation model; componentwise adaptive likelihood gradient algorithm; stage-coupled sentencing model; theft cases
Yujing LIU, Yifei JIN, Zhixin LIU, et al. Componentwise adaptive prediction theory under saturated observations and its application to stage-coupled sentencing. Sci Sin Inform, 2026, 56(9): 2283-2309, doi: 10.1360/SSI-2026-0235
数字化社会治理的复杂性机理与智能方法专题
强化学习驱动的有害信息智能检测方法
闫硕, 张兴伟, 张淳杰, 郑晓龙
中国科学: 信息科学, 2026, 56(9): 2310-2326
关键词 数字化社会治理; 有害信息检测; 强化学习; 大语言模型; 反事实奖励; digitalized social governance; harmful information detection; reinforcement learning; large language models; counterfactual reward
Shuo YAN, Xingwei ZHANG, Chunjie ZHANG, et al. A reinforcement learning-driven method for intelligent harmful information detection. Sci Sin Inform, 2026, 56(9): 2310-2326, doi: 10.1360/SSI-2026-0335
