Probabilistic modeling of human responses in complex interaction

该项目旨在开发计算能力，以可靠地检测并因此对隐含的用户偏好采取行动。它的目标是开发结合先进的非侵入性传感器测量交互任务中意识和非意识行为的技术，以实现对复杂数据和设计环境中人类目标寻找的非常高水平的计算机化支持。它计划使用用户的生理和偏好评估来捕捉他们与他们查看的数据的复杂交互，使用概率模型来积累信息以确定他们潜在的偏好，并提取关系以找到可能的“隐藏变量”，这些变量可能有助于解释和利用用户的选择。

The project aims to develop computational ability to reliably detect and hence act on implicit user preferences. It aims to develop techniques combining advanced non-intrusive sensor measures of conscious and non-conscious behaviour during interaction tasks to enable very high-level computerised support for human goal-seeking in complex data and design environments. It plans to use a user’s physiology and preference evaluation to capture their complex interaction with the data they view, probability models to accumulate information to identify their underlying preferences and extract relationships to find possible ‘hidden variables’ which may help explain and leverage the user's choices.