Human Computation Games for Discovering Social Media Content

Principal Investigator (PI): Assoc Prof Dion Goh

Co-Principal Investigator (Co-PI): Assoc Prof Alton Chua, Professor Lim Ee Peng

Collaborators: Mr Low Guanghao

Start Date: Nov 2014

End Date: On-going

Abstract: User-generated social media content has many uses but its proliferation has exacerbated the information overload problem, making it necessary for techniques to manage and hence facilitate their retrieval. One promising approach to content management is via crowdsourcing through gaming, where computer games serve as motivators for users to contribute their brainpower to address the problem. That is, while users are entertained through gameplay, crowdsourcing tasks are performed as a byproduct. We term such games as human computation games (HCGs), and aim to answer the question: How can games be leveraged for social media content discovery? Our objectives are thus:

  • Investigate design elements for effective HCGs
  • Understand HCG usage behavior
  • Examine the viability of different game genres for HCGs
  • Evaluate HCGs in terms of perceptions and performance

We adopt an interdisciplinary orientation involving understanding individual differences and their influence on game preferences and behaviors, as well as designing engaging and useful HCGs.

The research comprises two inter-related streams. Stream 1 involves constructing a framework for modeling individuals’ behaviors, perceptions and needs arising from HCG gameplay. Stream 2 first develops an inventory of HCGs to document their characteristics, followed by the development of a suite of HCGs for quality assurance, metadata generation, sentiment analysis and relatedness discovery, activities which are critical for social media content discovery.

This proposal is innovative in the following areas. First, we advance an interdisciplinary research program that emphasizes the role of the social sciences in HCG design. Next, unlike prior research that focuses on casual games, the project raises the profile of HCG research by investigating unexplored genres. Finally, there is a potential commercialization potential through a system for social media content discovery. This will be of interest to organizations that use social media to promote and/or improve their products and services.