基于智能Agent的远程学习者情感与认知识别模型——眼动追踪与表情识别技术支持下的耦合

情感与认知状态的准确识别是实现远程学习者与教学Agent有效互动的基础。只有有效识别出学习者的情感与认知状态,教学Agent在改变学习者行为态度、帮助学习者获取和理解知识、支持学习者认知发展方面才能取得预期的效果。现有的教学Agent普遍存在两方面问题:一是缺乏情感交互性,容易使远程学习者产生厌倦情绪;二是认知推断功能薄弱,对学习效果的促进作用不稳定。这主要是由于Agent对学习者状态的识别不够充分造成的。已有的学习者状态识别方法虽然在学习者情绪状态的识别方面各有优势,但却无法同时检测学习者的视域、学习情绪与认知状态。学习者的眼动追踪数据是判断学习者实时状态的重要指标,也是学习者与Agent进行情感交互的重要依据。结合表情识别和眼动追踪技术构建的基于智能Agent的远程学习者情感与认知识别模型,将眼动追踪与表情监控迭代识别、情感与认知识别过程相耦合,以提高远程学习者状态的识别准确率,改进Agent对学习者的情感和认知支持,为智能教学Agent与远程学习者交互机制的研究提供新的思路和方法。

An Emotional and Cognitive Recognition Model for Distance learners Based on Intelligent Agent--- The Coupling of Eye .Tracking and Expression Recognition Techniques

The accurate recognition of emotional and cognitive status is a significant foundation for effective interaction between instructional agents and distance learners. Only when learners＇ emotional and cognitive status were successfully identified, can instructional agents can achieve desired effects on changing learners＇behaviors and attitudes, helping learners to acquire and understand knowledge and support learners＇ cognitive development. There are two common problems in instructional agents nowadays ： One is the lack of emotional interaction, which is likely to lead to boredom among distance learners; another is the weak cognitive inference, whose promotion on learning effect is not stable. Both are caused by the inadequate identification of learners＇ status. The existing tech- niques on learners＇ status recognition have their own advantages, but they cannot detect learners＇ vision, emotion and cognitive status simultaneously. Learner~ eye tracking data is an important indicator in judging learner＇ s real - time status, and it is also an indispensable base of interaction between learners and the instructional agent. This study attempts to establish an Intelligent Agent- based Emotional and Cognitive Recognition Model for Distance learners by means of coupling facial expression recognition and eye tracking techniques, as well as the emotional and cognitive recognition processes, in order to raise recognition accuracy, improve emotional and cognitive supports for distance learners, and provide new ideas and methodology for similar studies.