Deep Bayesian Trust: A Dominant and Fair Incentive Mechanism for Crowd

Authors: Naman Goel, Boi Faltings1996-2003

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through numerical experiments, we show the robustness of our mechanism under various reporting strategies of the workers. In a preliminary study conducted on Amazon Mechanical Turk, we observe that the mechanism helps in eliciting effort and improving the quality of responses.
Researcher Affiliation Academia Naman Goel, Boi Faltings Artificial Intelligence Laboratory Ecole Polytechnique F ed erale de Lausanne Lausanne, Switzerland, 1015 {naman.goel, boi.faltings}@epfl.ch
Pseudocode Yes Mechanism 1 : The Deep Bayesian Trust Mechanism 1. Assign a set of tasks to the oracle o and obtain its answers on the tasks. 2. Initialize an Informative Answer Pool (IAP) with the answers given by oracle. [...]
Open Source Code No The supplementary material for this paper is available on authors website.
Open Datasets No The paper mentions 'Numerical Simulations' where 'The proficiency matrices of different workers were generated independently'. This implies simulated data, not a publicly available dataset. It also mentions 'a preliminary study conducted on Amazon Mechanical Turk' but provides no access details for any data from this study.
Dataset Splits No The paper describes simulation settings ('Workers were simulated to be hired in 4 rounds, with 5, 25, 125 and 625 workers in successive rounds.') but does not specify any training, validation, or test dataset splits for real-world datasets.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU, CPU models, memory) used for running the numerical simulations or the Amazon Mechanical Turk study.
Software Dependencies No The paper does not specify any software dependencies or their version numbers used for the experiments or simulations.
Experiment Setup Yes Workers were simulated to be hired in 4 rounds, with 5, 25, 125 and 625 workers in successive rounds. K was set to 2 in all the experiments discussed in the paper. The proficiency matrices of different workers were generated independently such that the diagonal entries Ai[g, g] g [K] were β(5, 1) distributed.