Computational models for intention handling

这项研究关注的是扩大计算认知理论的范围，以理解人类的执行控制功能，或者可以被视为“人类操作系统”。一个基本的问题是，人类的大脑是如何设定一个指向未来的意图的：一个决定在未来的时间执行一些行动（比如记得在一天结束的时候买牛奶），以及随后的执行活动，如果以及当环境暗示条件合适时。意图在智能计算理论中的作用是引导注意力和推理资源。本研究发展了认知理论，基于计算机模拟和实验室实验，关于意图是如何被人类认知结构所表征和操纵的。这对于理解人类如何同时执行多个任务具有实际意义。理解支持未来导向意图的认知微观结构与各种认知工程工作相关，例如，对于支持人类在动态环境中的表现的系统设计，在动态环境中，任务被中断，然后必须在一段适当的时间后或当环境发生变化时恢复。被打断的任务可以被看作是未来恢复活动的直接意图，了解哪些因素会影响这种意图是否可能被记住并成功执行是很重要的。

The research is concerned with expanding the scope of computational cognitive theories to understand the human executive control function, or what might be viewed as the 'human operating system.' One basic question is how the human brain sets a future-directed intention: a decision to execute some action at a future time (such as remembering to buy milk at the end of the day) and the subsequent execution that activity, if and when environmental cues signal that conditions are appropriate to do so. The role of intentions in computational theories of intelligence is to direct attention and reasoning resources. This research develops cognitive theories, based on computer simulations and laboratory experiments, about how intentions are represented and manipulated by the human cognitive architecture. This has practical implications for understanding how humans can do several tasks concurrently. Understanding the cognitive microstructure that supports future-directed intentions is relevant to various cognitive engineering efforts, e.g., for the design of systems that support human performance in dynamic environments in which tasks are interrupted and then must be resumed after some appropriate time or when the environment changes. Interrupted tasks can be viewed as a future directed intention to resume an activity,and it is important to understand what kinds of factors impact whether or not such an intention is likely to be remembered and executed successfully.