∗Cyber Security Group, Delft University of Technology, Delft, NetherlandsDecember 31st, 2021
FiledDecember 31st, 2021An example scenario in real-life would be crowdsourcing services like Amazon’s Mechanical Turk (MTurk). In the machine learning community, researchers use MTurk for manual data annotation or labeling for applications like behavior signal processing, computer vision and so on [10], [11], [12]. Here, annotators need to be ensured that they will receive a payment for their work, and researchers need to be ensured on the quality and correctness of these annotations, which is a challenging problem [12]. Note that the researchers (receiver) cannot trust the hash of annotations given by the annotators (sender) and use Amazon (TTP) for only escrowing purposes. Thus, the receiver either requests (i) the TTP to validate the correctness of the goods (if a TTP has the expertise), and the description is correctly calculated, or (ii) a sample of the goods from the sender as a commitment that will