Collaborative Research: From User Reviews to User-Centered Generative Design: Automated Methods for Augmented Designer Performance

协作研究：从用户评论到以用户为中心的生成设计：增强设计师性能的自动化方法

• 批准号: 2050130
• 负责人: Lu Wang
• 金额: $ 20.39万
• 依托单位: Regents of the University of Michigan - Ann Arbor
• 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: 2021-08-01 至 2024-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2050130&HistoricalAwards=false
• 关键词: Collaborative; Research; User; Reviews; Centered

This project investigates design processes where the unmet needs of users are elicited from social media, online forums, and e-commerce platforms, and translated into new concept recommendations for designers using artificial intelligence (AI). The motivation stems from the growing abundance of user-generated feedback and a lack of advanced computational methods for drawing useful design knowledge and insights from that data. The research will establish a rigorous computational foundation that (1) enables large-scale elicitation of user needs from online reviews using advanced natural language processing (NLP) algorithms, and (2) translates the elicited needs into the visual and functional aspects of new concepts using novel generative adversarial networks (GAN) algorithms. The theoretical innovations will advance the fundamental understanding of how AI can augment the performance and creativity of designers in early-stage product development processes. This project will boost national competitiveness in innovation by creating tacit opportunities for designing innovative, inclusive, and competitive products. The convergent research team will create outreach initiatives for STEM students, teachers, and underrepresented minorities, and engage with industry and research stakeholders to ensure technology-market fit and successful dissemination.The overarching goal of this project is to establish a transformative, data-driven paradigm for empathetic design that augments the ability of designers to uncover and address the critical yet latent needs of users at scale. The project will create scalable and computationally efficient NLP algorithms that capture the needs of ordinary users from reviews, identify the underlying usage contexts, and infer extreme use-cases to facilitate latent need elicitation. Focus groups and interviews involving ninety design experts and crowdsourced evaluators will be conducted to test the first research hypothesis: The NLP algorithms elicit needs that are nonobvious, difficult to identify, and provide significant value and originality. The project will build novel GAN architectures and algorithms for generative design of form and function conditioned on the elicited latent user needs. New multimodal deep regression models will be developed to evaluate the quality of the generated samples based on user feedback on existing products. Laboratory studies involving fifty subjects and fifty evaluators will be performed to test the second research hypothesis: The GAN-generated design recommendations significantly improve the quality and variety of the design concepts generated by human designers. The project will lead to broad societal outcomes by fostering designer-AI co-creation and innovation centered on empathy with users to bridge the gap between user need discovery and design outcomes.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.

该项目研究了从社交媒体、在线论坛和电子商务平台引出用户未满足的需求的设计流程，并利用人工智能（AI）转化为设计师的新概念建议。其动机源于用户生成的反馈日益丰富，以及缺乏先进的计算方法来从这些数据中提取有用的设计知识和见解。该研究将建立严格的计算基础，（1）能够使用先进的自然语言处理（NLP）算法从在线评论中大规模获取用户需求，以及（2）将所获取的需求转化为新概念的视觉和功能方面使用新颖的生成对抗网络（GAN）算法。理论创新将增进对人工智能如何增强设计师在早期产品开发过程中的表现和创造力的基本理解。该项目将通过创造设计创新、包容、有竞争力的产品的默契机会，提升国家创新竞争力。融合研究团队将为 STEM 学生、教师和代表性不足的少数群体制定推广计划，并与行业和研究利益相关者合作，以确保技术与市场的契合和成功传播。该项目的总体目标是建立一个变革性的、数据驱动的移情设计的范例，增强了设计师大规模发现和解决用户关键但潜在需求的能力。该项目将创建可扩展且计算高效的 NLP 算法，从评论中捕获普通用户的需求，识别潜在的使用上下文，并推断极端用例以促进潜在需求的挖掘。将进行由 90 名设计专家和众包评估人员参与的焦点小组和访谈，以检验第一个研究假设：NLP 算法引出不明显、难以识别的需求，并提供显着的价值和原创性。该项目将构建新颖的 GAN 架构和算法，用于根据引发的潜在用户需求进行形式和功能的生成设计。将开发新的多模态深度回归模型，以根据用户对现有产品的反馈来评估生成样本的质量。将进行涉及 50 名受试者和 50 名评估员的实验室研究来测试第二个研究假设：GAN 生成的设计建议显着提高了人类设计师生成的设计概念的质量和多样性。该项目将通过促进设计师与人工智能的共同创造和创新，以与用户的同理心为中心，弥合用户需求发现和设计成果之间的差距，从而产生广泛的社会成果。该奖项反映了 NSF 的法定使命，并通过评估被认为值得支持利用基金会的智力优势和更广泛的影响审查标准。

项目成果

Generative Aspect-Based Sentiment Analysis with Contrastive Learning and Expressive Structure

具有对比学习和表达结构的基于生成方面的情感分析

• DOI: 10.48550/arxiv.2211.07743
• 发表时间: 2022-11-14
• 期刊: ArXiv
• 作者: Joseph Peper;Lu Wang
• 通讯作者: Lu Wang