Supporting Designers in Learning to Co-create with AI for Complex Computational Design Tasks

支持设计师学习与人工智能共同创造复杂的计算设计任务

• 批准号: 2118924
• 负责人: Nikolas Martelaro
• 金额: $ 85万
• 依托单位: Carnegie-Mellon University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-08-15 至 2024-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2118924&HistoricalAwards=false
• 关键词: Supporting; Designers; Learning; Co; create

As modern design tasks grow increasingly challenging, artificially intelligent (AI) design tools have the potential to provide new support to designers and engineers. This project will work to enable a future of computational co-creation, in which humans and AI collaborate and learn from each other to create new designs. The project addresses a vital national need: to prepare the emerging workforce in design, engineering, and manufacturing to better solve the complex problems of today by collaborating with AI design tools. In practice, co-creation with AI presents a significant learning curve for designers. The research team will study how people learn to collaborate with AI on real-world design tasks. The team will continuously build and test new ways for AI and designers to learn and interact through conversational and graphical interfaces. Success in this project is expected to advance our understanding of how people learn to collaborate with AI on complex computational co-creation tasks. This project is expected to lead to new training techniques and software design guidelines for AI-enabled design tools and other human-AI co-creative tasks. The research team will share their new interface designs and strategies to support other researchers in studying human-AI co-creation and to support companies in developing new AI-enabled design tools. The team will also incorporate the developed tools and research knowledge into classes for university and high school students, introducing them to human-AI collaboration and preparing them to work with and develop such systems in the future.The research team will conduct iterative, human-centered design research to advance our understanding of how people learn to collaborate with AI on complex design tasks, while using widely available AI design tools, in the context of designing actively transforming structures. This is a complex, emerging manufacturing task. Steps in the research include: (1) Conduct a series of think-aloud activities to investigate how designers (try to) learn to collaborate with currently available AI design tools. The findings from these activities are expected to provide understandings of current challenges in human-AI design collaboration and will surface the strategies and mental models that people use when learning to collaborate with AI. (2) Prototype novel interface features to advance human–AI co-creation and learning. These will include a range of interactions and interface modalities, building upon theories of effective conversational exchange and supporting controlling actions, delegating actions, and negotiating goals and means. (3) Evaluate these interfaces and interactions to see how well they support designers in learning to interact and productively work with the AI. (4) Use a mix of conversation analysis and multimodal observations to understand how the interface prototypes influence human-AI co-creation. The research is expected to produce new approaches to help researchers study and design human-AI co-creation, including: (a) measures for assessing learning and collaboration in the context of human-AI co-creative tasks; (b) prototyping methods for human-AI collaborative systems; (c) interaction design guidelines for onboarding and supporting designers in human-AI co-creation; and (d) new theory about how conversational interfaces can support designers in learning to co-create with AI.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

随着现代设计任务变得越来越具有挑战性，人工智能（AI）设计工具有可能为设计师和工程师提供新的支持。该项目将致力于实现计算共同创造的未来，人类和人工智能相互合作，相互学习，创造新的设计。该项目解决了一个至关重要的国家需求：通过与人工智能设计工具合作，为设计、工程和制造领域的新兴劳动力做好准备，以更好地解决当今的复杂问题。在实践中，与人工智能的共同创造为设计师提供了一个重要的学习曲线。研究团队将研究人们如何学习在现实世界的设计任务中与人工智能合作。该团队将不断构建和测试人工智能和设计师通过对话和图形界面进行学习和互动的新方法。该项目的成功有望促进我们对人们如何学习在复杂的计算共同创造任务中与AI合作的理解。该项目预计将为人工智能设计工具和其他人类-人工智能共同创造任务带来新的培训技术和软件设计指南。研究团队将分享他们的新界面设计和策略，以支持其他研究人员研究人类与人工智能的共同创造，并支持公司开发新的人工智能设计工具。该团队还将把开发的工具和研究知识融入大学和高中学生的课程中，向他们介绍人类-人工智能协作，并为他们将来使用和开发此类系统做好准备。研究团队将进行迭代的，以人为本的设计研究，以促进我们对人们如何学习在复杂的设计任务中与人工智能协作的理解，同时使用广泛使用的人工智能设计工具，in the context上下文of designing设计actively积极transforming变换structures结构.这是一项复杂的新兴制造任务。研究的步骤包括：（1）进行一系列有声思维活动，以研究设计师如何（尝试）学习与当前可用的人工智能设计工具协作。这些活动的结果预计将提供对人类与人工智能设计合作当前挑战的理解，并将揭示人们在学习与人工智能合作时使用的策略和心理模型。(2)原型新颖的界面功能，以促进人类与AI的共同创造和学习。这些将包括一系列的互动和接口模式，建立在有效的对话交流和支持控制行动，授权行动，谈判的目标和手段的理论。(3)评估这些界面和交互，看看它们在多大程度上支持设计师学习与AI交互和高效工作。(4)使用对话分析和多模态观察的组合来了解界面原型如何影响人类-AI共同创造。该研究预计将产生新的方法来帮助研究人员研究和设计人类-人工智能共同创造，包括：（a）评估人类-人工智能共同创造任务背景下的学习和协作的措施;（B）人类-人工智能协作系统的原型方法;（c）人类-人工智能共同创造中的入职和支持设计师的交互设计指南;（d）人类-人工智能共同创造的方法。以及（d）关于对话界面如何支持设计师学习与人工智能共同创造的新理论。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估而被认为值得支持。

项目成果

Team Learning as a Lens for Designing Human-AI Co-Creative Systems

团队学习作为设计人类与人工智能共同创造系统的镜头

• DOI: 10.48550/arxiv.2207.02996
• 发表时间: 2022
• 期刊: ACM CHI 2022 Workshop on Generative AI and HCI
• 作者: Gmeiner, Frederic;Holstein, Kenneth;Martelaro, Nikolas
• 通讯作者: Martelaro, Nikolas

Physically Situated Tools for Exploring a Grain Space in Computational Machine Knitting

用于探索计算机针织中颗粒空间的物理定位工具

• DOI: 10.1145/3544548.3581434
• 发表时间: 2023
• 期刊: Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems
• 作者: Albaugh, Lea;Hudson, Scott E;Yao, Lining
• 通讯作者: Yao, Lining

An Augmented Knitting Machine for Operational Assistance and Guided Improvisation

• DOI: 10.1145/3544548.3581549
• 发表时间: 2023-04
• 期刊: Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems
• 作者: Lea Albaugh;S. Hudson;L. Yao

Exploring Challenges and Opportunities to Support Designers in Learning to Co-create with AI-based Manufacturing Design Tools

探索支持设计师学习使用基于人工智能的制造设计工具共同创造的挑战和机遇

• DOI: 10.1145/3544548.3580999
• 发表时间: 2023
• 期刊: CHI '23: Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems
• 作者: Gmeiner, Frederic;Yang, Humphrey;Yao, Lining;Holstein, Kenneth;Martelaro, Nikolas
• 通讯作者: Martelaro, Nikolas

FlexTure: Designing Configurable and Dynamic Surface Features

• DOI: 10.1145/3563657.3595995
• 发表时间: 2023-07
• 期刊: Proceedings of the 2023 ACM Designing Interactive Systems Conference
• 作者: Tate Johnson;Dinesh K. Patel;Humphrey Yang;U. S. Civici;Adriane Fernandes Minori;L. Yao