FEATURED VIDEO

Sponsored By:


SLIDE SHOWS
A Cisco commissioned study questioned 2,000 IT professionals and employees about risky behavior that could result in corporate data leaks. From losing laptops to letting strangers wander around their building, this is what was found.
Check out these new and improved enterprise IT management tools from CA.
The best multi-function printers (MFP), also known as all-in-ones, offer a slew of great features designed to streamline production and reduce costs. Here are a few of the market's best and brightest to come out recently.
INSIDE CHANNELWEB
techcareers logo Search Jobs:


  

Post Resume|Employers

Recent Post:


Automotive Market Segment Director
Silicon Labs seeking Automotive Market Segment Director in Austin, TX
spacer

BLOGS
The Channel Wire
June 20, 2008
Launched in October of 2006, The Neflix Prize is a contest that allows open competition among developers to help fine tune the company's algorithm. The competition centers on Cinematch, Netflix's internally developed movie recommendation system. The goal for entrants into the competition is to improve on the results Cinematch gives users by 10 percent.

Acknowledging that literature exists describing better ways to come up with movie matches for customers, Neflix has decided to put its money where its mouth is.

According to the Netflix Web site, "There are a lot of interesting alternative approaches to how Cinematch works that we haven't tried. Some are described in the literature, some aren't. We're curious whether any of these can beat Cinematch by making better predictions."

In order to beat Cinematch, the company outlines the criteria for contestants to beat.

"We suspect the 10% improvement is pretty tough, but we also think there is a good chance it can be achieved," Netflix writes on the contest Website. All participants are given a test set of data to test their algorithms on in order to see how closely the ratings predicted match actual ratings.

Netflix realizes that coming up with an algorithm that beats theirs will take time. The Netflix Contest opened on October 2, 2006 and runs until at least October 2, 2011. This gives contestants who got in on the ground floor a full five years to develop and refine their movie matching methods.

Each year there are two awards that developers are eligible for. The $1 million grand prize is always on the table, but in years where no one reaches the 10 percent goal, the Progress Prize is available and pays $50,000.

The winners of the Progress Prize are chosen by Netflix. "It goes to the team whose system we judge shows the most improvement over the previous year's best accuracy bar on the same qualifying test set. No improvement, no prize."

Perhaps the most important element of the Netflix Challenge centers on the source code of the algorithm. After registering and qualifying, developers have to be prepared to hand their home grown code to Netflix for their use.

The legalese that developers who want to claim a prize must submit to is extensive, but not necessarily draconian.

According to the official rules, "each Participant will be required to grant to Netflix an irrevocable, royalty free, fully paid up, worldwide non-exclusive license under the Participants' copyrights, patents or other intellectual property rights to use, review, assess, test and otherwise analyze the submissions (prediction sets, algorithm descriptions, and source code) and all their content in connection with this Contest."

Posted by Brian Kraemer at 1:52 PM
CHANNELWEB MARKETSPACE >> (Sponsored Links)
ADVERTISEMENT




CHANNEL SERVICES >>