Facilitating Deliberation with Real-time Crowd Workers For Better Prediction on Car Accident

less than 1 minute read

Instantaneous crowdsourcing, which prefetches possible future states and receiving crowd task results ahead the critical events really happens, introduced much potential to solve real-world problems that require accelerated human capabilities. However, real-world is complex, and there can be a magnificent amount of possible future states. Though humans are capable of predicting critical events, because uncertainties in predicting the future, their predictions can be either diverging or wrong. Crowdsourcing research has been shown the effectiveness of collective deliberation in such problems, but it can be limited in problems that only allows limited time for workers to do the task. We introduce efficient and quick collaboration techniques that can induce crowd workers to deliberate and do tasks better in real-time. We test our technique in the context of predicting car accidents.