Consumer Computing Lab

ACM Transactions on Knowledge Discovery from Data (TKDD)
Volume 6,  Issue 2   July 2012
Sanmay Das, Malik Magdon-Ismail
http://dl.acm.org/tipsvc.cfm?id=2297458&code=%27%2A%3C3%2AQ%3CC%272%40%20%20%0A&tick=D6F962ED2A53E623CBF86BD91D9E5255

“Collaborative media such as wikis have become enormously successful venues for information creation. Articles accrue information through the asynchronous editing of users who arrive both seeking information and possibly able to contribute information. Most articles stabilize to high-quality, trusted sources of information representing the collective wisdom of all the users who edited the article. We propose a model for information growth which relies on two main observations: (i) as an article’s quality improves, it attracts visitors at a faster rate (a rich-get-richer phenomenon); and, simultaneously, (ii) the chances that a new visitor will improve the article drops (there is only so much that can be said about a particular topic). Our model is able to reproduce many features of the edit dynamics observed on Wikipedia; in particular, it captures the observed rise in the edit rate, followed by 1/t decay. Despite differences in the media, we also document similar features in the comment rates for a segment of the LiveJournal blogosphere”.

From MIT News (08/24/12) by Larry Hardesty

“Massachusetts Institute of Technology (MIT) researchers have developed Qurk, a database system that automatically crowdsources tasks that are difficult or impossible to perform computationally. For example, images in a Qurk database could be sorted according to the approximate age of the people depicted, the appeal of the depicted locations as travel destinations, or any other attribute whose assessment would require human judgment. “You can just say, ‘I have this collection of images, and I want to sort them by how cute they are,’ and the system will actually figure out how to implement a sort over your data set,” says MIT’s Adam Marcus. The researchers found that, while ranking provided more accurate sorting, Qurk fared better and was much less expensive than other systems. Marcus notes that Qurk also will enable users to specify a group of attributes that might be useful for pre-filtering. The system will then evaluate those attributes on the fly, determining which, if any, actually increase the efficiency of the join operation”.

From http://www.technologyreview.com/news/429118/artificial-intelligence-powered-by-many-humans/#

“University of Rochester researchers have developed Chorus, an approach to virtual personal assistants that creates a smart artificial chat partner from small contributions from many crowdsourced workers. During testing, Chorus was asked for travel advice and the system showed that it could be smarter than any one individual in the crowd, because multiple people were contributing to its responses. “It shows how a crowd-powered system that is relatively simple can do something that [artificial intelligence] has struggled to do for decades,” says Rochester professor Jeffrey Bigham. The researchers aimed to find a new way to increase the power of crowdsourcing, which is traditionally limited to simple, isolated tasks. “What we’re really interested in is when a crowd as a collective can do better than even a high-quality individual,” Bigham says. As part of the Chorus system, any new chat updates from the user are passed along to many crowd workers, who are asked to suggest a reply. The suggestions are then voted on by crowd workers to determine which one will be sent back to the user. The final step creates a working memory that ensures that Chorus’ replies reflect the history of the conversation.”

CEO of VIPnet has visited our laboratory on 18.th April 2012 and was informed of the activities going on in CCL laboratory.