算法治理

当前，随着以“算力新基建、数据新要素、AI大模型”为特征的算力经济时代来临，以chatGPT为代表的AIGC应用掀起人工智能新一轮浪潮，全球人工智能发展逐步从“探索期”向“成长期”过渡在技术和产业上进入重要转型阶段。在此背景下人工智能发展和数据、算法安全问题深度交织融合影响用户隐私、公民权益、商业秘密、国家安全等各个方面，AIGC的创新应用和数据安全治理工作也进入全新的阶段。本报告重点聚焦并梳理了人工智能应用中较为独特或更突出的安全问题，全面分析了当前全球人工智能数据安全和算法治理的主要现状和最新动态，基于全球相关治理实践和我国实际情况，构建了AIGC数据安全和算法治理框架。

随着新型电力系统建设的加速推进，电网投资的力度持续攀升，企业须探索更加合理高效的电网投资策略，以实现最优的综合效益。为此，从经济、社会、环保和安全的维度出发，以电网项目投资的综合效益为优化目标，对投资组合策略进行研究。首先，应用Shapley值法对各效益函数的占比进行分摊，以呈现不同效益指标的特性。其次，在考虑投资能力、负荷需求等关键约束的基础上，设计了电网项目投资组合优化模型，并采用全局和声搜索算法进行求解。最后，构建算例对投资组合策略进行验证。算例结果表明，所提方法能够协助决策者在新形势下制定最优的电网投资策略。 With the accelerated advancement of the construction of new-type power systems and the increasing intensity of grid investment, power grid enterprises have to explore more rational and efficient investment strategies to achieve optimal comprehensive benefits. To this end, taking into account the economic, social, environmental, and safety dimensions and optimizing the comprehensive benefits of grid project investments, a study on investment portfolio strategies is conducted. Firstly, the Shapley value method is applied to allocate the proportions of various benefit functions, revealing the characteristics of different benefit indicators. Secondly, given the key constraints such as investment capacity and load demand, an optimal model for grid project investment portfolios is developed, and the global harmony search algorithm(GHSA) is employed for solution finding. Finally, a case study is constructed to investigate the investment portfolio strategy. The results demonstrate that the proposed approach can assist decision-makers in formulating optimal investment strategies in the new situation.