Crowdsourcing Wins
for Netflix

Had enough of “crowdsourcing”? That’s the term for the much-hyped trend of asking the general public to help solve complex problems. Historically, busi-nesses tackled tricky challenges by assigning their best people to confront and solve them. Later, they turned to pricey consulting companies. Today, executives throw the problem to the general public by announcing it through Twitter or a website, then offering a prize to whoever comes up with the optimal solution.

Crowdsourcing can be effective and probably costs less than hiring a consultant. But does it deliver results quick enough? Not necessarily.

The Netflix Prize has become the
poster child for crowdsourcing. The
movie-rental company wanted to
improve the efficiency of the algorithms
it uses to recommend movies to its
customers. The contest, which began
October 2, 2006, promised a $1 million
grand prize to whoever could deliver a
prediction algorithm that performed
at least 10 percent better than the
then-current Netflix algorithm, called
Cinematch. Netflix knew it might take
time, announcing the contest would run
“through at least October 2, 2011.”
The grand prize was finally awarded
last September to a team called BellKor’s
Pragmatic Chaos for a solution it sub-
mitted the previous July. The team con-
sisted of Bob Bell and Chris Volinsky
from the Statistics Research group in
AT&T Labs and Yehuda Koren, who
had recently left AT& T Labs for Yahoo!
Research in Israel. The winning algo-
rithm, said Netflix, delivered a 10.06
percent improvement over Cinematch.

Netflix makes no secret of the solution, called “The BellKor Solution to the Netflix Grand Prize”; the 10-page paper on how it works is available to all.

How should it follow such success? Saying another contest will be announced soon, Neil Hung, Netflix chief product officer, said, “While the first contest has been remarkable, we think Netflix Prize 2 will be more challenging, more fun, and even more useful to the field.”

Is It Plagiarism or Just a Copy?

On the Internet, nobody knows you’re a dog… but with the proper search engine, everybody can know if you’ve plagiarized someone else’s work. Whether it’s a high-school book review, a speech given to a local city council, or a song lyric, search engines can help determine if a writer has copied someone else’s work wholesale or even merely lifted a catchy phrase without attribution.

Just because catching plagiarism is
so easy doesn’t mean the plagiarist will
be caught. If nobody checks, nobody
knows. Increasingly, though, many
people do check: teachers, political ene-
mies, editors. And now there are the pla-
giarism cops, the informal term used for
the increasing number of websites that
monitor the Internet for stolen words
and works, according to Diane Mapes
at MSNBC.com “In the last few months
alone,” she wrote, “the p-word has been
tossed—rightly or wrongly—at New
York Times columnist Maureen Dowd,
Twilight author Stephenie Meyer, a
consortium of Bollywood movie produc-
ers, and some sperm researchers from
Newcastle University in England.”
Why do we plagiarize? Sometimes it’s
intentional. Why write a paper or speech

if we can just copy one instead? However, it’s often unintentional, some experts say. Perhaps we jot down a catchy phrase, not realizing our subconscious dredged it up from something we read months ago.

The plagiarism cops offer several
services, says Mapes. For example,
a company called iParadigms LLC
offers one it calls Turnitin Originality
Checking, describing it this way:
“Originality Checking allows educators
to check students’ work for improper
citation or potential plagiarism by com-
paring it against continuously updated
databases. Every Originality Report pro-
vides instructors with the opportunity
to teach their students proper citation
methods as well as to safeguard their
students’ academic integrity.”
Another is Copyscape, a service
provider with a Spartan user interface
that uncannily resembles the home page
of the Google search engine. Type the
URL of your document in the search
window and in seconds Copyscape tells
us if it’s found pages with the same con-
tent. The service is designed primarily
to help authors and publishers defend
their works by detecting unauthorized
duplication on the Internet and teachers
detect plagiarism by students; in fact,
my son is required to use it for his high-
school English and history class essays.
Copyscape could also be used by anyone
to learn if anything has been plagiarized
from any Web-based sources. 

 

Alan Zeichick ( [e-mail]) is a senior member of the ACM and technology analyst in the San Francisco Bay Area; follow him at twitter.com/zeichick.

DOI: 10.1145/1655737.1655739