vish
New Member
Posts: 17 
|
Post by vish on Apr 26, 2016 5:34:44 GMT
A similar project was done by DiSalvo in 2008/09. Where they tried to address some of the local issues and uses by utilising the robotics sensing tools. The project goal was to involve community users to use and explore robotics and sensing device for addressing local community issues. The experiments conducted to test air pollution and noise pollution showed that there were significant interests from the community users to address the issues. Although the project goal was achieved with a great success, there were few concerns that were noticed. However, to have a participatory design for a project, time will act as a major factor, and may not be the best solution for every community issues. The Tripwire project clarifies the point on having community involvement. However, this project had no rigid time constraint and made a good case in displaying the community users’ effort and novelty in designing a robotics and sensing tool. Talking about the user involvement, there were some good number of users who were involved in this project, but the authors also mentioned about those community users who just appreciated the work verbally but had zero contribution towards participation. Project similar to this may need a guidance from people who have technical and scientific research background. So, unless a funding agent or company is willing to interest themselves in such projects, the project may be hindered.
|
|
|
|
Post by rushil on Apr 26, 2016 5:51:00 GMT
This somewhat reminded me of Michael Bernstein's talk last semester. In his work, he basically was using crowdworkers for solving problems. If I remember correctly, one of the bigger challenges was that the efforts were often misguided which resulted in "temporary leaders" within the crowd. It was not explicitly stated, but the leaders assumed the role of breaking down the task and guiding the efforts.
I believe as the crowd gets bigger, the efforts gets spread out which adds to chaos. A little guidance (nudge) is needed to ensure constant progress.
|
|
|
|
Post by judithodili on Apr 26, 2016 8:17:16 GMT
Re: Three tasks used in this paper for evaluating unsupervised crowd planning failed because of starvation or derailment in the recursive workflow generating process. Do you think non-expert turkers can ever reliably design workflow for complex tasks? Why or Why not? I doubt that this is possible. Planning workflows like this require a lot of experience, as mentioned by Nathan and Judy. People who have been through the workflow many times will probably come up with more efficient ways to do things, such as programmers who customize their text editors or anyone who has pushed back on a process suggested by HR. However, I doubt that most crowd workers (at least on Turk) spend enough time doing any entire workflow to come up with an optimal workflow. The requesters probably don't know terribly well either, since they are not completing the task, but at the very least they have the 100-foot view and the statistics to investigate inefficiencies in the system. Is it a good idea to investigate? Yeah, but I'm not holding out for much except a "Suggest feedback" button on task design. + 1 This idea sounds great in theory, but I honestly don't think Turk workers put that much thought into their tasks. From the perspective of a researcher/requester, this makes perfect sense - allow the turker the opportunity to produce the highest quality work. However, it makes the assumption that requesters and workers have the same goals of producing quality work vs doing as many tasks as possible to make some money. If such a system were to work, my intuition is that it has to be implemented formally by the platform (e.g Amazon Turk) or the tasks has to be broken down manually by the requester, and managed with workflows. |
|
|
|
Post by mrivera on Apr 29, 2016 14:55:00 GMT
To Brandon's point "We don't think 'type T,H,E' now, but we did when we were first learning to type. I think part of the difficulty with crowd work is the inability to adapt tasks as workers become more familiar with tasks. You can't easily build expecting a learning curve b/c you can't count on the same workers.": Is there any research or stats on whether turk workers choose to do similar tasks or do they go for money over everything else?
|
|
