Programming digital computers to perform complicated concurrent tasks usually executed by several human beings is important in business intelligence decisions, but not easy. How should multifaceted or interconnected tasks be automated on computers to schedule and monitor human workers in timely and cost-efficient ways? Barowy et al. succinctly present automatic algorithms for the effective scheduling, pricing, and quality control of concurrent activities that humans perform to accomplish a task, such as the recognition of alternative license plates at a tollbooth.
Can a single programming system be automatically and effectively used to schedule employees who work in parallel to achieve quality tasks? The authors present a domain-specific language called AutoMan for coalescing and making transparent the efficient schedules of designed human and digital computation activities for accomplishing alternative, complex real-world tasks.
AutoMan provides an easy interface that allows users to specify efficient functions for incorporating crowdsourcing into computations. Users can create functions to define the nature of questions and responses for human workers in AutoMan. The programmer can specify the desirable degree of accuracy of the computational results of all human intelligence tasks in the AutoMan environment. The authors succinctly present algorithms for effectively advertising jobs with rewards for workers to accept the duties on time, and for certifying the quality of job allocations.
The authors clearly and convincingly present evidence to illustrate the effective use of AutoMan in applications such as the automatic recognition of license plates. Without a doubt, this paper will promote new research in the areas of the Internet of Things (IoT) and digital computation. I strongly encourage all current and future software engineers to read and contribute to the unique, insightful ideas presented.