普适计算环境中用户意图推理的Bayes方法  

Bayesian method for intent prediction in pervasive computing environments

作　　者：易鑫 喻纯[1,4,2,3] 史元春[1,4,2,3] Xin YI;Chun YU;Yuanchun SHI(Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China;Global Innovation exchange Institute, Tsinghua University, Beijing 100084, China;Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing 100084, China;Key Laboratory of Pervasive Computing, Ministry of Education, Tsinghua University, Beijing 100084, China)

机构地区：[1]清华大学计算机科学与技术系,北京100084 [2]清华大学北京信息科学与技术国家研究中心,北京100084 [3]清华大学普适计算(教育部)重点实验室,北京100084 [4]清华大学全球创新学院,北京100084

出　　处：《中国科学：信息科学》2018年第4期419-432,共14页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:61521002;61672314;61572276);清华大学科研基金(批准号:20151080408);网络多媒体北京市重点实验室资助项目

摘　　要：本文阐述了通过Bayes方法来预测用户交互意图的建模方法过程和推理过程.在自然交互界面上,用户不再是严格地通过离散明确的交互操作完成交互,而是通过连续、非确定的多模态数据表达交互意图.在解释用户的交互意图时,既可以使用"黑盒子"的机器学习方法,也可以利用"白盒子"的基于用户行为建模的方法.后者中的用户建模,其本质是通过计算的方法来刻画用户的行为能力,对于理解用户意图和探索自然交互的计算原理具有重要的科学意义.文章回顾了近年来人机交互研究中主要采用的智能算法,向读者厘清不同方法之间的差别,并通过我们实验室的具体研究工作展示用户建模的方法和Bayes推理的建模方法过程和推理过程.This paper describes the principle and examples of predicting users＇ intent using the Bayesian method. In natural interfaces, users use continuous, undetermined multi-modal data to express their interaction intention rather than using discrete, determined actions. In order to interpret their interaction intend, we can either use ＂black-box＂ machine-learning methods or use ＂white-box＂ user-behavior-modelling methods. The essence of the latter is to computationally model the users＇ interaction ability, which is crucial to understanding the users and exploring the computation principle of natural interaction. This paper summarizes the popular intelligent algorithms used in recent HCI researches, discusses methods of user modelling and the Bayesian method the difference between these methods, and illustrates the using some works from our laboratory.

关 键 词：BAYES方法 机器学习 意图推理 用户建模 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]