Human and Machine Collaboration in Creative Design

Establishing that machines cannot be creative in the same way as humans, we propose a computational model which allows human and machine to collaborate on creative design in a social structure similar to human-human collaboration. We then discuss specific architectural problems associated with the design of such an interactive, collaborative, and intelligent system. Design is a complex cognitive process which involves creative skills, artistic intuitions, as well as a rich repertory of knowledge. Further- more design is not description of what is, it is exploration of what might be (Mitchell (8)). According to our experiences with designers, exploration is further characterized by being a process, driven by the desire to change an imperfection, that pushes many limits in the open world. A creative design is one that meets a particular goal and at the same time compromises a set of constraints and current technological limitations. To free machines from the predefined world or to enrich them with interactions from nature (for a more detailed article, see (14)), we must allow humans to help them in a similar way that humans benefit from the calculation and data processing power of machi- nes. In this paper we propose an approach which integrates several previously built intelligent design systems in an open and distribu- ted architecture. Most importantly, this architecture will facilitate human-machine collaboration and human and machine creativity to co-exist and co-contribute to the creative design process. An open architecture contradicts the traditional algorithmic nature of software and organizes its functionalities at the c ognitive level of task distri- bution instead of the control level of task performance. We propose to use the method of software agents to design and develop such an architecture as depicted in Figure 1. Automatic and semi-automatic agents collaborate with human designers, and they cooperate with each other seamlesslyto achieve a given goal. Some agents are refle- ctions (mirrors) of the cognitive inner workings of a person while others are computational because they take care of detailed design works such as constraint processing, search, and data visualization. Many of the agents of this architecture have been developed such as the case-based reasoning, constraint processing, and multimedia browsing units. Here we discuss the challenging issues we encounte- red in the development of this system: 1) what contributions human and machine can each provide to the creative design process, 2) what are the mirroring and processing roles of agents, and 3) what will be the collaborative technique which allows contributions from humans and machines to take place in the creative design process.