Stochastic models for the optimal fusion of social and sensory information in transparent interactions (C01)

透明交互中社会和感官信息最佳融合的随机模型 (C01)

• 批准号: 496442018
• 金额: --
• 依托单位: Campus Institut für Dynamik biologischer Netzwerke (CIDBN)
• 依托单位国家: 德国
• 项目类别: Collaborative Research Centres
• 资助国家: 德国
• 项目状态: 未结题
• 来源: https://gepris.dfg.de/gepris/projekt/496442018?language=en
• 关键词: Stochastic; models; optimal; fusion; social

Activities such as joint play, hunting or foraging involve the use of shared information (e.g., about actions of others), and require the representation and integration of social and sensory information to guide the agents’ actions. We will establish a formal theory and analysis framework for dyadic continuous perceptual report tasks employed in this CRC. We will construct models for the formation and optimal integration of representations for sensory and social information in a pair of mutually interacting agents, construct methods for the data-driven inference of inter-agent coupling and behavioural dynamics in the presence of temporal and inter-individual variability, and theoretically predict optimal reporting strategies to optimise expected reward for cooperative vs. competitive settings.

共同玩耍、狩猎或觅食等活动涉及使用共享信息（例如，关于他人的行为），并需要社会和感官信息的表示和整合来指导智能体的行为。我们将建立一个正式的理论和分析框架的二元连续知觉报告任务中使用的CRC。我们将构建模型，用于在一对相互作用的代理中形成和最佳整合感官和社会信息的表示，构建方法，用于在存在时间和个体间变异的情况下对代理间耦合和行为动力学进行数据驱动的推断，并从理论上预测最佳报告策略，以优化合作与竞争环境的预期奖励。