Crowdscanner: Viral social media challenges can be solved 'within the hour'

Researchers reckon that they can solve the US government sponsored crowdsourcing game, Tag Challenge 'within the first hour of the competition' if viral communication by social media works as predicted.
Written by Eileen Brown, Contributor

Researchers at the University of Southampton reckon that they can solve the US government sponsored crowdsourcing game, Tag Challenge ‘within the first hour of the competition’ if the viral effect of communicating via social media works as well as predicted.

In order to win the Tag challenge, you need to have a team of social communicators located in five locations around the world who are prepared to share information in order to claim the prize.

A crowdsourcing team is being assembled to solve the puzzle. It is called “CrowdScanner”, and anyone can join the group. The site is being built in collaboration with experts from University of Southampton and University of California San Diego.

Dr. Victor Naroditskiy, researcher in the Agents, Interaction and Complexity group at the University of Southampton is the European Team leader of Crowd Scanner. The team will use the data they gather to learn what makes ‘mobilization through social networks successful’.

“Scientists have long conjectured theoretical conditions under which something "goes viral". However, this question is far from solved, especially in the context of time-critical tasks involving a large geographical area. Is success dependent on some special "hubs": people capable of recruiting large numbers? Is it dependent on the nature of the incentives people have or is it something else altogether”?

The team plans to write a research article describing the data and reflecting on the challenge and add this to on-going work towards a general Web platform for large-scale ad-hoc team building.

Naroditskiy reckons that CrowdScanner will have a better chance for success because it will enable the formation of a very large ad-hoc team, involving people who have never met before. He explains:

‘If a group of friends try to work collaboratively, their numbers and geographical spread are likely to be limited. First of all, real-world social networks exhibit high "clustering", which means that most of my friends' friends are also my own friends".

"This clustering means that the team will likely be stuck within specific social groups, rather than spread more widely. Moreover, scientists have also observed that friendship also exhibits high geographical clustering".

"This means that my friends and friends' friends are more likely to live in the same area as me, rather than be spread over multiple cities or areas with significantly different demographics within the same city".

The ad hoc nature of social media could mean that it might be difficult to structure a team based on existing social relationships, no matter where they are located. In contrast, CrowdScanner provides a platform for people to create a very widespread team by joining their clusters into a larger mesh.

CrowdScanner also has a different incentive mechanism. It does not only reward people who find the target, but also the people who recruited them.

As in the DARPA Red Balloon challenge, this is a powerful way to quickly increase the size of the team.

The financial incentive is structured slightly differently too. The entire reward will be given to the participants who will actually do the work. Dr. Manuel Cebrian, member of the Crowdscanner team explains:

” If you submit a photograph of a target to CrowdScanner and they win, they will give you $500. In addition, if you invite a friend, and your friend submits a photograph, they get $500, but you also get $100. To help spread the word, the recruiters of the first 2000 members will get $1 for each person they invite to join the team.

“We are not interested in the money; we just want to learn more about the power of social media in information gathering  tasks like this one,” Dr. Cebrian added.

If this works, there are positive implications for crowdsourcing for community good.  We rush to try to help good causes, we retweet information about missing cats, children and bags.

I wonder whether, in this instance, people will participate purely for the financial rewards, skewing the results of the research. Or will our natural desire to help out and share information, overcome our jaded attitude to social engagement?

Related content:

Editorial standards