Modeling Task Complexity in Crowdsourcing

Introduction

This is the companion page for HCOMP2016 submission, titled "Modeling Task Complexity in Crowdsourcing".
This page presents a full description of the feature classes we designed for task complexity, and the optimisation method for MFLR. 


The 3 classes of task features for predicting complexity (metadata, semantic, and visual).




Optimization method for MFLR




For the details of Karush-Kuhn-Tucker complementary slackness conditions, please refer to: 
  [1] Stephen Boyd and Lieven Vandenberghe. Convex optimization. Cambridge university press, 2009. 

The derivation from Equation 6 to 7 is similar to the one in:   
  [2] Chris Ding, Tao Li, Wei Peng, and Haesun Park. Orthogonal nonnegative matrix t-factorizations for clustering. In KDD’06, pages 126–135. ACM, 2006.