Today, I think I finally picked out my Master’s project, with the help of Dr. Michael Shafae. I will be working towards the Netflix $1 Million prize. Yeah, I am probably late given that the contest went live in 2006, but a winning solution (one that beats the Netflix current system by 10%) hasn’t been found yet.
This raises an interesting question of what is going to happen if or when someone comes up with the answer while I am in the middle of my research. On the one hand, I didn’t win the grand prize, but on the other hand I shouldn’t be doing this for the money anyway, but rather for the sake of knowledge.
Luckily, Dr. Shafae suggested that we concurrently work on visualization techniques for the data and publish a few papers relating to the visualization techniques specifically (of course he would suggest that, he’s a graphics professor to begin with!) and, more generally, publish the data mining techniques I will have explored.
Plus, there’s absolutely no reason I can’t try to improve on the techniques beyond the 10% mark (if I ever even get that close).
However, if I ever do reach the 10% mark, and if I reach it within the lifespan of my graduate career, I will be kicking myself forever and ever for not starting it earlier and eventually receiving the $1M.
By the way, the whole reason I went into this (besides, obviously, the prize and a feeling that I have a shot against these BellKor guys), is because Dr. Shafae and I engaged in a conversation about Singular Value Decomposition and its applications in the real world. I was just curious what it was since it was in the textbook and that’s how the whole thing got started.
Anyway, it’s good to finally have a professor on my side since I am really good at starting projects and never finishing them. I am sure he will force me to show progress every week, which is definitely a good thing. In true professor/mentor fashion, he has already shown me how to keep my references organized using Bibdesk (think of it as painting fences, washing cars, and sanding decks).