IdeaHound: improving large-scale collaborative ideation with crowd-powered real-time semantic ...
summary
IdeaHound combines a shared whiteboard where participants drag ideas into clusters with real-time aggregation to build a low‑dimensional idea map using pairwise co-clustering probabilities fed to a dimensionality reduction (t-SNE). An integrated crowdsourcing workflow captures semantic judgments from 200+ Mechanical Turk contributors as they ideate, avoiding 1,000 outsourced triplet tasks. The system yields maps that correlate better with human similarity judgments, preserves idea quantity/creativity/diversity, and raises but then reduces mental effort.
AI-written summary · about
chapters (7)
- 0:00The challenge of crowd ideas
- 1:14Introducing IdeaHouse
- 4:17Semantic mapping concepts
- 7:02Integrating tedious tasks
- 8:37How IdeaHouse works
- 13:40Evaluation and results
- 17:51Conclusion and Q&A
