A-Life sculpture: eliciting complex interactions

联合项目摘要新媒体艺术家Nell Tenhaaf将与计算机科学家Melanie Baljko博士合作创建公共互动场景。在这些场景中，人类将与一个或多个人工代理交互，每个人工代理的实施例由电子元件(例如发光二极管(LED)阵列、2声道音频显示器)组成。像这样的低保真实施例是为该项目刻意选择的；它们用于描述使用抽象形式的代理，并在比参与与动画、卡通般的角色互动的层面上更基本的层面上与互动者互动。我们计划设计一种机制，通过这种机制，人类互动者会发现种群中的一个人工代理人实际上代表了他或她自己。此外，这个“交互-代表”代理调解人类用户与种群中其他代理的通信。人类群体中的所有智能体都将使用多种发音模式：人类将使用触摸、语音、接近和动觉以及文本等模式；智能体将使用“手势”模式(由其多色灯光的像素化显示提供)和声音发音模式。出现的互动结构将是互动者集体活动的产物，因此他们各自都是动态艺术品的共同创造者。这部作品的新颖性和创新性将在于引发的交际过程的结构；这种结构将以主体和人类互动者之间的适应和会话话轮为特征，这种结构的启发是一种艺术努力。此外，这种结构将支持这样的假设，即是主体的行为，而不是其体现，触发了人类用户对人工主体的交际意图的归因，这一发现与人机交互研究高度相关。Tenhaaf和Baljko将在他们之前工作的基础上，共同开发构建互动场景的计算技术、响应用户输入的算法、代理体系结构以及让人们参与交流过程的视觉和其他感官策略。

Joint Project SummaryNew Media artist Nell Tenhaaf will work with computer scientist Dr. Melanie Baljko to create public interactive scenarios. In these scenarios, humans will interact with one or more artificial agents, each of whose embodiment consists of electronic components (e.g. arrays of light-emitting diodes (LEDs), 2-channel audio displays). Low-fidelity embodiments such as these have been deliberately selected for the project; they serve to depict agents using abstracted form and to engage interactants at a level that is more conceptually fundamental than that which is engaged in interactions with animated, cartoon-like characters. We plan to devise mechanisms whereby the human interactant will discover that one of the artificial agents in the population actually is representative of him- or herself. Further, this "interactant-representative" agent mediates the human user's communication with other agents in the population. All agents in the population will make use of multiple modes of articulation: the human will employ modes such as touch, voice, proximity and kinesthetics, and text; the agents will employ the mode of "gesture" (afforded by its pixellated display of multicoloured lights) and the mode of sound articulation. The interactive structure that emerges will be the product of the collective activity of the interactants, so that they each function as co-creator of a dynamic artwork. The novelty and innovation of this work will lie in the structure of the elicited communicative processes; the structure will feature adaptation of the agents and the human interactant to one another and conversational turn-taking, the elicitation of such structure is an artistic endeavour. In addition, such structure will support the hypothesis that it is the behaviour of an agent, and not its embodiment, that serves to trigger the attribution of communicative intention by human users to artificial agents, a finding that is highly relevant to human-computer interaction research. Tenhaaf and Baljko, building upon their previous work, will jointly develop the computational techniques for building the interactive scenarios, the algorithms for responding to the user's input, the agent architecture, and the visual and other sensory strategies for engaging people in communication processes.