Accession Number:

AD1051572

Title:

Optimizing Human Input in Social Network Analysis

Descriptive Note:

Technical Report,13 Jul 2016,12 Apr 2017

Corporate Author:

University of Texas at Austin Austin United States

Personal Author(s):

• Shakkottai,Sanjay

Report Date:

2018-01-23

Pagination or Media Count:

67.0

Abstract:

The study focused on developing new bandit algorithms for online optimization e.g. matching tasks to human agents. The attached technical reports provide details on the formulations and results. Specifically, the technical reports focused on a backlog minimization formulation for matching tasks and agents, as well as a contextual bandit formulation focusing on dimensionality reduction.

Descriptors:

• Social Networks
• RANDOM variables
• machine learning
• information systems
• probability
• stochastic processes
• information theory
• generative models
• algorithms
• computational science
• phase transformations
• artificial intelligence
• dimensionality reduction

Subject Categories:

• Information Science

Distribution Statement:

APPROVED FOR PUBLIC RELEASE