What is it about?
Bayesian social learning refers to situations in which agent learn from observing the actions of other agents. In this work we consider the impact of "noisy observations" in a basic model for such a setting. Agents sequentially make a binary decision based on their observations of previous agents and on their own binary valued private signal. A classic result in such settings is that information cascades may occur in which agents eventually ignore their private signals and simply follow the crowd. In our model with observation noise, information cascade still occur, but the probability of an incorrect cascade is not monotonic in the noise level, i.e. sometimes more noise is better.
Featured Image
Photo by Chandler Chen on Unsplash
Why is it important?
This provides new and counter-intuitive insights into the behavior of social learning models in the presence of observation noise.
Read the Original
This page is a summary of: Information Cascades With Noise, IEEE Transactions on Signal and Information Processing over Networks, June 2017, Institute of Electrical & Electronics Engineers (IEEE),
DOI: 10.1109/tsipn.2017.2682798.
You can read the full text:
Contributors
The following have contributed to this page
