Apr 17, 2012, 11:25:25 PM4/17/12
Online labor markets, such as Amazonâ€™s Mechanical Turk, have been
used to crowdsource simple, short tasks like image labeling and
transcription. However, expert knowledge is often lacking in such
markets, making it impossible to complete certain classes of tasks. In
this work we introduce an alternative mechanism for crowdsourcing
tasks that require specialized knowledge or skill: communitysourcing â
€” the use of physical kiosks to elicit work from specific
populations. We investigate the potential of communitysourcing by
designing, implementing and evaluating Umati: the communitysourcing
vending machine. Umati allows users to earn credits by performing
tasks using a touchscreen attached to the machine. Physical rewards
(in this case, snacks) are dispensed through traditional vending
mechanics. We evaluated whether communitysourcing can accomplish
expert work by using Umati to grade Computer Science exams. We placed
Umati in a university Computer Science building, targeting students
with grading tasks for snacks. Over one week, 328 unique users (302 of
whom were students) completed 7771 tasks (7240 by students). 80% of
users had never participated in a crowdsourcing market before. We
found that Umati was able to grade exams with 2% higher accuracy (at
the same price) or at 33% lower cost (at equivalent accuracy) than
traditional single-expert grading. Mechanical Turk workers had no
success grading the same exams. These results indicate that
communitysourcing can successfully elicit high-quality expert work
from specific communities.
I will also spend some time talking about our LocalGround project.
LocalGround allows students and community members to use paper maps
for collecting local geo-spatial knowledge. Users annotate paper maps
using colored markers and symbols. These annotations are automatically
extracted and visualized with other data sources and forms of media.
Local Ground was used by teenagers from Richmond, California for
planning of a public park, and by Oakland youth to document healthy
food zones in their communities.
Bio: Tapan Parikh is an Assistant Professor at the School of
Information at the University of California, Berkeley. Tapan's
research interests include human-computer interaction (HCI), mobile
computing, paper and voice UIs and information systems for
microfinance, smallholder agriculture, global health and education.
For more then 10 years, Tapan has been designing, developing and
deploying information systems for communities - initially in India,
and now around the world. Tapan and his students have started several
technology companies serving community-based organizations (CBOs), non-
governmental organizations (NGOs), governments and non-profits. He
holds a Sc.B. degree in Molecular Modeling with Honors from Brown
University, and M.S. and Ph.D. degrees in Computer Science from the
University of Washington, where his dissertation won the William Chan
Memorial award. Tapan has also received the NSF CAREER award, was
named TR35 Humanitarian of the Year in 2007, and has won best paper
awards at several HCI and CS conferences.