|
Activity Number:
|
364
- Network Science: Statistical Approaches and Beyond
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 12, 2021 : 12:00 PM to 1:50 PM
|
|
Sponsor:
|
WNAR
|
|
Abstract #316877
|
|
|
Title:
|
Social Learning: Degree-Weighted Updating and Convergence Speed
|
|
Author(s):
|
Yiqing Xing* and Xin Tong and Xiao Han and Yusheng Wu
|
|
Companies:
|
Johns Hopkins University and University of Southern California and University of Science and Technology of China and University of Southern California
|
|
Keywords:
|
Social learning;
DeGroot model;
Convergence speed;
Heterogeneous learning weights;
Stochastic block model
|
|
Abstract:
|
People tend to rely more on important neighbors — how does this pattern affect social learning outcomes? We propose a degree-weighted DeGroot social learning model: each agent forms belief by averaging the beliefs of her neighbors in the previous period, with more weights put on the signals from more central neighbors — those with larger degrees. We capture the intensity of this pattern by a single parameter, allowing for general functional forms subject to mild regularity conditions. We show when the intensity increases, the social learning process converges faster. We prove this result both for fixed networks as well as random networks generated by stochastic block models.
|
Authors who are presenting talks have a * after their name.
