I-Corps L: TheDesignExchange- Nuturing a National Design Innovation Ecosystem

通过国家科学基金会创新团队学习计划(i-Corps L)，该项目将把设计交换转变为一个高影响力的可持续发展的企业，培育一个全国性的设计创新生态系统。以人为中心的设计正在成为创新和产品开发的重要途径。虽然有许多项目将以人为中心的设计原则介绍给感兴趣的个人，但从这些项目到成为专家的进展缺乏支持。对于那些专注于以人为本的过程的研究部分的公司来说尤其如此，这些公司设计并与潜在客户和用户一起进行研究，为创新和产品开发工作提供信息。设计交流旨在支持个人在设计掌握水平上取得进步，同时鼓励开发进行设计研究和应用以人为本的设计方法的最佳实践。设计交流将使设计方法与设计问题相匹配，以帮助设计师和设计研究人员就何时以及如何应用设计方法做出明智的决定，并促进从业者之间在交换专业支持方面的讨论。另一个目标是利用从设计交流中获得的知识来确定和教授与各种方法相关的基本技能，为新近毕业的学生提供更流畅的理解，了解如何、为什么以及何时应用特定的设计研究方法。设计交流中心有潜力产生广泛的影响，为设计界和那些有兴趣加入这些领域的人提供有关创新和设计过程的信息来源，同时作为社区讨论的在线聚会场所。设计交换团队通过对82个设计过程模型和300多种设计方法的深入分析，创建了设计方法分类。正在进行的工作分为四类：(1)通过与设计从业者举办的一系列讲习班完善分类；(2)收集有关设计过程案例研究，分析设计方法的使用；(3)利用机器学习算法开发设计方法推荐系统；(4)开发互动和智能门户网站。通过分析方法之间是否频繁共现，发现基于方法协方差的预测比基于问题内容的预测具有更高的查准率；通过对方法协方差数据进行谱聚类，方法可以被自动划分为专家给定的分组，准确率为92%。

Through the NSF Innovations Corps for Learning Program (I-Corps L), this project will transform TheDesignExchange into a high impact sustainable enterprise, nurturing a national design innovation ecosystem. Human-centered design is emerging as an important approach to innovation and product development. Although there are many programs to introduce the principles of human-centered design to interested individuals, progressing from these programs to becoming an expert is poorly supported. This is especially true for those that focus on the research portion of the human-centered process, those that design and conduct research studies with potential customers and users that inform innovation and product development efforts. TheDesignExchange aims to support individuals as they progress through the levels of design mastery, while encouraging the development of best practices for conducting design research and applying human-centered design methods. TheDesignExchange will match design methods to design problems in order to help designers and design researchers to make informed decisions about when and how to apply methods, and facilitates discussions between practitioners in an exchange of professional support. Another goal is to leverage knowledge gained from theDesignExchange to identify and teach the underlying skills associated with the various methods, providing recent graduates a more fluid understanding of how, why, and when to apply specific design research methods. TheDesignExchange has the potential to produce broad impacts as a source of information on innovation and the design process for the design community, and for those interested in joining these fields, while acting as an online meeting place for community discussions. TheDesignExchange team has created a taxonomy of design methods collected from an in-depth analysis of 82 design process models, and the collection of over 300 design methods. Ongoing work falls into four categories: (1) refinement of the taxonomy through a series of workshops held with design practitioners; (2) collection of relevant design process case studies analyzed for the use of design methods; and (3) development of a design method recommendation system, using machine learning algorithms; (4) development of an interactive and intelligent web portal. By analyzing whether methods frequently co-occur with one another or not, it was found that predictions based on method covariance have higher precision-recall performance than predictions based on problem content; and by using spectral clustering on the method covariance data, methods can be automatically divided into expert-given groupings with 92% accuracy.