辅助社交训练严肃游戏中虚拟角色行为表现的不确定性模型

目的 建立行为可信的虚拟角色能够使严肃游戏更加有趣,提升用户使用的体验感。尽管严肃游戏的图形渲染技术已经日趋成熟,但现有的虚拟角色行为表现方式多采用确定性模型,很难反映虚拟角色行为表现的多样性。方法 本文构建了符合辅助社交训练的严肃游戏剧情,采用智能体来描述虚拟角色,赋予虚拟角色视觉、听觉双通道感知。基于马斯洛动机理论,采用食物、休息、交流和安全等动机来描述情绪的产生,利用大五（OCEAN）个性模型来描述虚拟角色的不同个性差别。用外部刺激和内部动机需求来计算情绪强度,利用行为树描述虚拟角色的行为。运用正态云模型处理虚拟角色行为表现的不确定性,并以行走方向、社交距离、交流时身体朝向3个典型的行为表现给出了具体处理方法。结果 在所实现的游戏原型系统中,对于虚拟角色的自主行为和行为表现的不确定性进行了用户体验测试。结果表明,在场景探索任务中,虚拟角色的自主行为模型能减少用户探索场景所耗费的时间,并且可以促进用户与虚拟角色交流;在行为表现测试中,本文模型的自然性评价要高于确定性模型。结论 本文虚拟角色行为模型在一定程度上可提升用户的体验感,有望为建立行为可信的虚拟角色提供一种新的途径。     

基于认知的情感多Agent交互建模研究

以多角色CVLE建模和角色情感为研究主线,在结合国内外虚拟环境、认知理论、情感理论、多Agent系统(Multi-Agent System,MAS)理论等相关文献基础上,针对CVLE具有多角色、开放式、协作性、感知性的特点,提出一种基于角色组织与感知的多Agent框架模型;针对虚拟角色认知与情感交互影响特点,提出了基于认知的情感交互建模,可实现具有开放式多角色的协同虚拟学习环境系统。     

虚拟角色不确定性行为建模研究

为了增强虚拟环境中虚拟角色行为表现的多样性和变化性,从虚拟角色不确定性行为的相关概念入手,系统地研究了虚拟角色的不确定性行为及其建模的基本问题．指出并证明了个体与群体不确定性行为的关系,基于不确定性函数证明了虚拟角色不确定性行为建模的3种途径一不确定性决策、规划和动作建模,分析了不确定性行为建模的具体实现方法．最后基于坦克群对战仿真实验,说明了相关概念的合理性,验证了相关结论的正确性,实验表明,虚拟角色不确定性行为对仿真应用具有重要的影响．     

基于PFA的虚拟角色行为辅助设计工具

在虚拟训练系统中,如何构建高度真实的虚拟角色是一个根本性的问题.不确定性是人类行为的一个重要特征,对其建模有助于增强虚拟角色的真实感和可信度.现有的基于规则、Petri网和FSM的行为建模工具无法满足不确定性行为建模的需求.在对虚拟角色行为不确定性进行定义的基础上,利用概率有限自动机(PFA)对不确定决策与规划进行建模,形成了行为概率自动机(BPFA).基于BPFA设计并实现了一个行为设计工具UBTool,方便了不确定性行为的设计与实现.最后以坦克作战行为为例展示了该工具的有效性.     

基于概率区间的不确定性动作建模研究

对不确定性动作建模有助于提高虚拟角色行为表现的逼真度和可信度。现有行为模型由于动作缺少必要的不确定性而显得重复、单调,为此,提出了利用概率区间对角色不确定性动作建模的方法。基于区间代数提出了概率区间的概念,并指出了概率区间的完备性问题,着重研究了正态概率区间的数学运算及其完备化的加性和乘性方法。与代数区间相比,概率区间能够更加精确地描述动作的不确定性效应。以坦克虚拟角色为例,研究了正态区间在不确定性动作建模中的应用,通过对射击动作的不确定性建模验证了所提方法的合理性和有效性。     

一种虚拟角色的情绪认知评价模型

虚拟角色在电子教育、动漫游戏和电子商务等领域具有广泛的用途,目前的虚拟角色情绪设计仍然处于手工阶段,需要花费大量的人力成本且大多数虚拟角色行为简单,不能令人信服。本文基于心理学理论和OCC理论,建立虚拟角色的认知结构,给出虚拟角色的动机形式化描述。虚拟角色的情绪状态可由情绪决策产生式规则来驱动,情绪强度可通过模糊推理系统来计算。实验结果表明该情绪模型可以增强教育软件的界面友好性,使虚拟角色在行为上更加人性化。     

基于弱化Agent的虚拟维修样机行为建模方法

为了建立具有行为和反应能力的虚拟维修样机行为模型,真正体现虚拟样机的模型特征,引入基于弱化Agent的建模方法对虚拟维修样机进行行为建模。给出了弱化Agent的特点和行为定义模式,研究了虚拟维修环境下行为模型的表示,行为模型的划分和行为模型的输出,给出了行为建模的具体实现过程,并把上述方法应用于某大型设备虚拟维修样机模型的行为建模中。通过应用实例可以看出应用此方法进行虚拟维修样机的行为建模极大的减少了虚拟维修样机与虚拟环境的交互数据量,可以很好的满足虚拟维修仿真实时性的应用要求。     

地理位置和时间感知的表示学习框架

现有时空感知的表示学习框架无法对强时空语义的实际场景存在的“When”、“Where”和“What”3个问题给出一个统一的解决方案。同时,现有的时间和空间建模上的研究方案也存在着一定的缺陷,无法在复杂的实际场景中取得最优的性能。为了解决这些问题,本文提出了一个统一的用户表示框架—GTRL (geography and time aware representation learning),可以同时在时间和空间的维度上对用户的历史行为轨迹进行联合建模。在时间建模上,GTRL采用函数式的时间编码以及连续时间和上下文感知的图注意力网络,在动态的用户行为图上灵活地捕获高阶的结构化时序信息。在空间建模上,GTRL采用了层级化的地理编码和深度历史轨迹建模模块高效地刻画了用户的地理位置偏好。GTRL设计了统一的联合优化方案,同时在交互预测、交互时间预测以及交互位置3个任务上进行模型学习。最后,本文在公开数据集和工业数据集上设计了大量的实验,分别验证了GTRL相较学术界基线模型的优势,以及在实际业务场景中的有效性。     

情绪和动机驱动的虚拟人自适应感知模型

通过引入注意力机制,以愉悦度、激活度、优势度(PAD维度)观测模型为理论基础,以OCC(Ortony-CloreColins)情绪模型为情绪选择范围,提出一种适用于3D动漫游戏的虚拟角色多通道自适应感知模型.该模型侧重于研究情绪状态对虚拟角色动态感知和动机的影响,考虑生理状况(体力)对感知的影响,通过行动、评估、行动周期模型驱动虚拟人的感知和动机,并以有限状态机实现虚拟人在不同需求下的情绪转换.最后设计并实现了一个基于本模型的虚拟人寻找目标物的原型系统.实验结果表明,文中模型能在较大程度上保持游戏的高可信性、趣味性和沉浸感.     

一个情感驱动的3D虚拟角色行为动画创作工具

情感作为人类具有智能的一个重要体现,是创建丰富细腻的虚拟智能体的不可或缺的环节。本文结合认知对情感的影响,提出一个改进的情感建模方法和一种新的行为建模方法,设计具体的情感情境,并通过行为树的行为组织形式对虚拟环境中的行为进行管理,以产生具体、真实的虚拟角色情感行为。提出＂个性-情感-情绪＂的三层情感层次结构,并最后完成了一个实验系统。实验结果证明,该情感模型和行为组织模型能有效地反映虚拟角色在虚拟环境中的智能情感行为。     

Computer agents versus avatars: Responses to interactive game characters controlled by a computer or other player

Computer and videogames often require that users interact with other characters on the screen that represent other real people or characters that are controlled by computer code running within the game. The difference between game play with other avatars (player-controlled characters) or agents (characters controlled by the computer) may influence the engagement a game player experiences. This study investigated the effects of agency (avatar versus agent) and the type of gaming activity (competition versus cooperation) on physiological arousal and subjective evaluation of play. A 2 (avatar, agent)×2 (competition, cooperation) within-subject experiment was conducted (N=32). Players exhibited greater physiological arousal to otherwise identical interactions when other characters were introduced as an avatar rather than an agent. Furthermore, the co-player's source of agency interacted with the type of gaming activity. The results have implications for understanding how different forms of representation in virtual worlds and games will affect psychological responses in the contexts of entertainment, learning and the conduct of serious work.     

Predicting students' knowledge after playing a serious game based on learning analytics data: A case study

Serious games have proven to be a powerful tool in education to engage, motivate, and help students learn. However, the change in student knowledge after playing games is usually measured with traditional (paper) prequestionnaires–postquestionnaires. We propose a combination of game learning analytics and data mining techniques to predict knowledge change based on in-game student interactions. We have tested this approach in a case study for which we have conducted preexperiments–postexperiments with 227 students playing a previously validated serious game on first aid techniques. We collected student interaction data while students played, using a game learning analytics infrastructure and the standard data format Experience API for Serious Games. After data collection, we developed and tested prediction models to determine whether knowledge, given as posttest results, can be accurately predicted. Additionally, we compared models both with and without pretest information to determine the importance of previous knowledge when predicting postgame knowledge. The high accuracy of the obtained prediction models suggests that serious games can be used not only to teach but also to measure knowledge acquisition after playing. This will simplify serious games application for educational settings and especially in the classroom easing teachers' evaluation tasks.     

Relations Between the Tendency to Invest in Virtual Presence,Actual Virtual Presence,and Learning Outcomes in Educational Computer Games

This study examines whether the effect of virtual presence on learning is influenced by learner characteristics. More specifically, the focus is on how the variation in the actual experience of virtual presence and learning is related to learners' individual tendency to invest in virtual presence for educational computer games. According to these tendencies, 59 eighth graders were divided into 2 groups (high vs. low tendency) and performed a computer-based learning game. Virtual presence and learning outcomes were measured by means of a questionnaire. Correlation and regression analyses were used, and a moderation model was structured to analyze the relationship between the variables. Whereas the tendency to invest in virtual presence does not moderate the positive interaction between virtual presence and learning, correlation and regression results between both groups show that a stronger tendency to invest in virtual presence increases virtual presence and both, trivial- and nontrivial learning outcomes were evinced. In the eyes of learners, interestingly, their tendency to invest in virtual presence for educational computer games was significantly lower than for conventional entertainment games.     

个体交互行为的平滑干预模型

基于交互行为的用户特征提取和身份认证方法是一种重要的身份识别方式,但高频用户的交互行为模式和操作习惯相对稳定,易被欺诈者模仿,使得现有模型对此类欺诈行为的误判较高.如何使得用户行为主动平滑变化且可区分,成为解决上述问题的关键.针对此问题,提出一种基于个体交互行为系统平滑干预模型:首先,根据用户历史交互行为日志从多个维度...     

Identification of trusted interactive behavior based on mouse behavior considering web User's emotions

Under existing network security technology, it is still possible for hackers to impersonate legitimate users and invade a system for malicious destruction. Therefore, this study constructs a user's unique mouse behavior pattern to identify a trusted interaction behavior in a real environment and quantify the effects of different emotions on mouse behavior and the accuracy of the user's trusted interaction behavior identification. First, mouse data was collected for 8 user's trusted interactions on an academic study website (AML). These data were used to construct the basic trusted interaction model by a big data analysis method called a random forest. Second, in a repeated measurement experiment, 18 participants completed tasks on the AML under different emotions, and the emotions' impact on the mouse behavior and accuracy of the user's trusted interaction identification was analyzed. In the results, the accuracy of the trusted interaction behavior identification based on mouse behavior reached 91.82%, and the error rate was lower than 8.18%. Significant differences were observed in horizontal velocity, velocity, and traveled distance under different emotions. However, there was no significant difference in the accuracy of a user's trusted interaction behavior identification under different emotions. Based on these results, the trusted interaction behavior of web users can be accurately identified based on the user's mouse behavior pattern. The user's mouse behavior differs under different emotions, but there is no significant difference on the identification of the user's trusted interaction behavior. The findings help to provide another protection layer for network information security.     

人机交互中的智能体情感计算研究

机器的情感是通过融入具有情感能力的智能体实现的,虽然目前在人机交互领域已经有大量研究成果,但有关智能体情感计算方面的研究尚处起步阶段,深入开展这项研究对推动人机交互领域的发展具有重要的科学和应用价值。本文通过检索Scopus数据库选择有代表性的文献,重点关注情感在智能体和用户之间的双向流动,分别从智能体对用户的情绪感知和对用户情绪调节的角度开展分析总结。首先梳理了用户情绪的识别方法,即通过用户的表情、语音、姿态、生理信号和文本信息等多通道信息分析用户的情绪状态,归纳了情绪识别中的一些机器学习方法。其次从用户体验角度分析具有情绪表现力的智能体对用户的影响,总结了智能体的情绪生成和表现技术,指出智能体除了通过表情之外,还可以通过注视、姿态、头部运动和手势等非言语动作来表现情绪。并且梳理了典型的智能体情绪架构,举例说明了强化学习在智能体情绪设计中的作用。同时为了验证模型的准确性,比较了已有的情感评估手段和评价指标。最后指出智能体情感计算急需解决的问题。通过对现有研究的总结,智能体情感计算研究是一个很有前景的研究方向,希望本文能够为深入开展相关研究提供借鉴。     

Probabilistic assessment of user's emotions in educational games

We present a probabilistic model to monitor a user's emotions and engagement during the interaction with educational games. We illustrate how our probabilistic model assesses affect by integrating evidence on both possible causes of the user's emotional arousal (i.e., the state of the interaction) and its effects (i.e., bodily expressions that are known to be influenced by emotional reactions). The probabilistic model relies on a Dynamic Decision Network to leverage any indirect evidence on the user's emotional state, in order to estimate this state and any other related variable in the model. This is crucial in a modeling task in which the available evidence usually varies with the user and with each particular interaction. The probabilistic model we present is to be used by decision theoretic pedagogical agents to generate interventions aimed at achieving the best tradeoff between a user's learning and engagement during the interaction with educational games.     

基于PageRank的多维度微博用户影响力度量

以新浪微博为研究对象,提出一种适用性更广、考虑因素更全面的微博用户影响力度量算法,将用户基本属性、用户交互行为和用户博文内容三个维度因素融入传统PageRank算法中,提出了一种多维度微博用户影响力度量算法——MDIR（multi-dimension influence rank）。实验结果表明,MDIR算法相较于其他常用的五种影响力度量算法,能更加全面、真实地反映微博用户的实际影响力。