Options
2020
Conference Paper
Titel
Learning a lightweight representation: First step towards automatic detection of multidimensional relationships between ideas
Abstract
Moving ideation from a closed paradigm (companies) to an open one (crowd) yields several benefits: (1) the crowd allows the generation of a large number of ideas and (2) its heterogeneity increases the potential in obtaining creative ideas. In practice, however, the crowd often fails at generating innovative solutions, leading to duplicate or ideas that use each other's description. Thus, it is practically and economically unfeasible to sift through this large number of ideas to select valuable ones. One promising solution to overcome this issue is identifying relationships between idea text descriptions, such as duplicate, generalize, disjoint, alternative solution. Existing approaches rely either on human judgment, which is expensive and requires domain experts or automatic approaches which compute similarity i.e. one dimension and do not consider other relations. To find complex relationships between idea texts, a first logical step is to fully structure ideas into logic-based representation. However, logic-based representations are very expensive to obtain since ideas' texts do not adhere to any specific structure. Our goal is then to come up with a representation that can be learned by the machine, while still being expressive to allow establishing relationships between ideas. This research in progress introduces an approach based on a sequence-to-sequence learning approach, which allows the machine to learn a lightweight structural representation that is used next to establishing multidimensional relationships between ideas (i.e. different kind of relations between ideas). Based on our investigation, we found out that ideas contain the following patterns: what the idea is about (e.g. window with heat-sensitive material), how it works (e.g. it lights up) and when it works (e.g. in case of fire). Those extracted patterns are then compared with the corresponding patterns of other ideas to establish relations. Our preliminary investigation shows promising results to learn such lightweight structural representation and leverages it in identifying complex relationships between ideas.