论文研究对话代理人工偏袒与关系持久性机制

1. 12:00
   arXiv · Human-Computer Interaction

   可信 AI 之爱：作为可工程化立场的人工偏袒与关系持续性

   论文提出对话智能体人工偏袒的行为机制，由 caring（在每用户保障服务底线之上分配有限交互盈余）与 particularity（在固定推理机器上从关系历史累积的每用户状态估计）两部分构成。

Current heat 9·Comparable peak 10（Oct 8）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.