Neural network-based ideation learning for intelligent agents: e-brainstorming with privacy preferences

Abstract

Creative thinking provide ways to many technical ideas, were new innovations will come to existence. For a single problem there can be numerous innovative solutions, it is very important to make evaluation and selection of ideas to achieve better results. E-brainstorming is an electronic version of sharing ideas. This paper integrates associative thinking of humans with an intelligent agent method to develop a neural network-based learning agent that can be represented as e-brainstorming session participant. Computer aided brainstorming decision model CABDM is built to construct an environment where neural network-based agents can learn rules from the database and make decisions. Additionally, privacy preferences, flexible ideation map construction are adopted to enhance e-brainstorming technique. The preliminary evaluation results indicates that the proposed work advances the existing agent-based e-Brainstorming by introducing neural network-based agents, with flexible ideation map to reduce production blocking problem by improving the productivity of e-brainstorming session.