Wednesday, September 23, 2009

Crowdsourcing finally snags the Netflix $1,000,000 prize.

It was inevitable. The competition that Netflix launched to improve its movie recommendation algorithm by 10% is finally over.

While the challenge took a while to crack the end was apparently tantalizingly close, with different cohorts posting their breakthroughs within minutes of each other.

Crowdsourcing is not new, but this recent application suggests an interesting lesson. While separate teams made strong progress in isolation, they reached a point beyond which they could not advance the algorithm further towards the goal. It took a final step of amalgamating the efforts of different groups and merging their algorithms together. One can only imagine the messy mathematical mash-up that entailed, but it resulted in the crucial incremental efficiency which enabled the 10% barrier to be breached.

It affirms how much more progress we can make through collaboration when conquering tough challenges. Several heads are better than one (as long as they’re the right heads!).

Full Wired magazine article here...

No comments: