An automated method for generating crowdsourcing design tasks based on knowledge graphs
Liu Minxia, Junde Lan, Lin Gong, Zhenchong Mo, Xinyou Yuan, Xirui Zhao · 2023
With the rapid development of Internet technology and sharing economy model, especially the popularity of mobile Internet has led to the outbreak of massive user requirement and various modes of personalized product requirement, so new design models such as crowdsourcing design and group wisdom collaborative design have emerged for massive personalized requirement. In this crowdsourcing context, the automated task generation for the design process is an important part to distinguish the new design model from traditional design, and is expected to solve the problem of lack of means to automate the design task generation, which is the core driving force for the downstream design stage such as resource matching of design platform. This paper is oriented to the product design process and mainly includes the following contents: this paper addresses the problem of automated generation of crowdsourcing design tasks, and carries out research on the construction of the crowdsourcing design requirements graph and the mapping of crowdsourcing design requirements to crowdsourcing design tasks. Based on the knowledge graph approach, Neo4j software is used to construct a multi-domain user requirement graphs, extract crowdsourcing design requirement entities as well as obtain word vectors based on Word2Vec model and perform cosine similarity calculation between entities to support the accurate mapping of user requirements to crowdsourcing design tasks and generate crowdsourcing design task packages. In order to integrate a crowdsourcing design task automated generation tool for crowdsourcing designers, a crowdsourcing design task automated generation method is proposed. Based on the crowdsourcing design task automation generation method proposed in this paper, the crowdsourcing design task automation generation tool is developed and validated by application examples, and the shortcomings of the study are discussed, and the outlook on the subsequent research directions is proposed.