Swissnoise: Online Polls with Game-Theoretic Incentives
Abstract
There is much interest in crowdsourcing information that is distributed among many individuals, such as the likelihood of future events, election outcomes, the quality of products, or the consequence of a decision. To obtain accurate outcomes, various game-theoretic incentive schemes have been proposed. However, only prediction markets have been tried in practice. In this paper, we describe an experimental platform, swissnoise, that compares prediction markets with peer prediction schemes developed in recent AI research. It shows that peer prediction schemes can achieve similar performance while being applicable to a much broader range of questions.
Cite
Text
Garcin and Faltings. "Swissnoise: Online Polls with Game-Theoretic Incentives." AAAI Conference on Artificial Intelligence, 2014. doi:10.1609/AAAI.V28I2.19031Markdown
[Garcin and Faltings. "Swissnoise: Online Polls with Game-Theoretic Incentives." AAAI Conference on Artificial Intelligence, 2014.](https://mlanthology.org/aaai/2014/garcin2014aaai-swissnoise/) doi:10.1609/AAAI.V28I2.19031BibTeX
@inproceedings{garcin2014aaai-swissnoise,
title = {{Swissnoise: Online Polls with Game-Theoretic Incentives}},
author = {Garcin, Florent and Faltings, Boi},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2014},
pages = {2972-2977},
doi = {10.1609/AAAI.V28I2.19031},
url = {https://mlanthology.org/aaai/2014/garcin2014aaai-swissnoise/}
}