Information Technology Reference
In-Depth Information
quality enough to obtain a refund for poor service. However, as long as such a large
coalition has not formed, a reward scheme based on peer prediction is sufficient to pun-
ish each individual user for deviating from truthful reporting, and can be realized at low
cost. Thus, a lying coalition would have to be created in a coordinated fashion, and such
coordinated action would be detectable by other means. This opens another wide range
of applications.
6
Conclusions
The internet has enabled wide distribution of user-contributed content whose correct-
ness cannot be verified. Much of this content is reported by agents with ulterior motives
and may often not reflect the truth. I have discussed ways of providing incentives to
agents to provide such content truthfully. I believe that such mechanisms are of fun-
damental importance for the future use of reputation forums, sensor nets and crowd-
sourcing applications on the internet. They also have other applications in multi-agent
systems, such as service monitoring.
While work so far has shown an interesting range of mechanisms to encourage truth-
ful reporting, many open questions remain. The biggest issue is clearly the dependence
on knowledge of prior probability distributions that are not always available. The opin-
ion poll mechanism we described is a first step but still has to be generalized to elicit
more complex information than just binary signals. Also, as it stands it has little protec-
tion against collusive behavior.
Another issue is how to provide rewards. Paying monetary rewards is often not prac-
tical, and one needs to experiment with other forms of rewards, such as reputation or
privileges that will be valued in similar ways as money.
Acknowledgements.
I thank Radu Jurca who has worked with me on this topic for
many years, and Karl Aberer for fruitful discussions on reputation and community
sensing.
This work has been supported in part by Opensense project (839-401) in the Nan-
otera.ch program.
References
1. Hu, N., Pavlou, P.A., Zhang, J.: Can online reviews reveal a product's true quality?: empirical
findings and analytical modeling of online word-of-mouth communication. In: Feigenbaum,
J., Chuang, J.C.I., Pennock, D.M. (eds.) ACM Conference on Electronic Commerce, pp.
324-330. ACM (2006)
2. Aberer, K., Sathe, S., Chakraborty, D., Martinoli, A., Barrenetxea, G., Faltings, B., Thiele,
L.: Opensense: Open community driven sensing of environment. In: ACM SIGSPATIAL
International Workshop on GeoStreaming, IWGS (2010)
3. Savage, L.J.: Elicitation of personal probabilities and expectations. Journal of the American
Statistical Association 66, 783-801 (1971)
4. Lambert, N.S., Shoham, Y.: Eliciting truthful answers to multiple-choice questions. In:
Chuang, J., Fortnow, L., Pu, P. (eds.) ACM Conference on Electronic Commerce, pp. 109-
118. ACM (2009)
