arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2409.02823cs.HC
收藏

设计矛盾:助力还是阻碍?

Design Contradictions: Help or Hindrance?

Aron E. Owen, Jonathan C. Roberts

AI总结:

本文探讨设计矛盾在AI驱动设计中的作用,指出LLM偏向相似性而创造力需发散性,并呼吁重新审视传统设计方法,探索利用生成式AI激发数据可视化创新的新途径。

摘要:

数据可视化领域对创新想法的需求推动我们探索新的创造性方法。将两个或多个创造性词汇组合在一起,尤其是那些相互矛盾的词汇,可以对创造过程产生积极影响,激发新颖的想法和设计。随着我们迈向人工智能驱动的设计,一个开放性问题随之出现:这些设计矛盾能否与人工智能工具产生积极作用?目前,答案是否定的。人工智能系统,如大型语言模型(LLMs),依赖的是倾向于产生相似性的算法,而创造力往往需要发散性和新颖性。本海报发起一场对话,探讨如何驱动人工智能系统更具创造力并生成新想法。这项研究邀请我们重新审视传统设计方法,并在人工智能驱动的世界中探索新途径。我们能否应用传统设计中使用的相同技术,例如双钻石模型,还是需要为设计工程引入新方法?我们如何利用生成式人工智能快速设计可视化并构思新想法?本文旨在开启这一关键对话,并就人工智能在推动数据可视化创造力方面的潜力提供实践性见解。

英文摘要:

The need for innovative ideas in data visualisation drives us to explore new creative approaches. Combining two or more creative words, particularly those that contradict each other, can positively impact the creative process, sparking novel ideas and designs. As we move towards AI-driven design, an open question arises: do these design contradictions work positively with AI tools? Currently, the answer is no. AI systems, like large language models (LLMs), rely on algorithms that engender similarity, whereas creativity often requires divergence and novelty. This poster initiates a conversation on how to drive AI systems to be more creative and generate new ideas. This research invites us to reconsider traditional design methods and explore new approaches in an AI-driven world. Can we apply the same techniques used in traditional design, like the double diamond model, or do we need new methods for design engineering? How can we quickly design visualisations and craft new ideas with generative AI? This paper seeks to start this critical conversation and offers practical insights into the potential of AI in driving creativity in data visualisation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑