Saturday, November 22, 2008

Aside: Netflix $1m competition on improving it's recommendation engine

If You Liked This, Sure to Love That  ( Free registration may be required).

The competition is really about coming up with a better collaborative filtering algorithm than Netflix's Cinematch. The discussion forum for the competition has open discussions on implementations and algorithms. 
A scientist & his team from AT&T lead the top 10 performance leaders that Netflix maintains on this website.

The book Programming Collective Intelligence by Toby Sebaran from O'Reilly has more technical background material on this.

Tuesday, November 18, 2008


Pongal (Tamil) - (lit.) to boil over; (fig.) an enthusiastic banter. 

I'll cover, over a long period of time, these topics. These are not just for interviews at Amazon/Google/Yahoo. Programming the Internet, and developing any non-trivial application require these. In fact, these are just the starting points. Next come, machine learning algorithms, approximation algos for hard problems, randomized algos etc.

References:
1) The Art of Computer Programming, Donald E. Knuth
As an original inventor teaching computer science, he walks you through the historical thought process that went into creating the various algorithms - the drawbacks of earlier algorithms and how that caused a subsequent development, and the intuition behind such development, and treats each topic holistically.

A good teacher develops the intuition - which helps in better application and further development.

2) Introduction to Algorithms, Cormen et al.
A big bazaar(all things available) book. The quality of each topic covered does'nt sulk, but lacks the continuity and a holistic treatment of Knuth.

3) Algorithm Design Manual, Steven S Skiena

The prose level is a muck. This is'nt a book to learn from. But the second part - a classified collection of problems, with references to implementations available online, gets you kickstarted.

Related URLs:

4) Algorithms on Strings, Trees and Sequences: Computer Science and Computational Biology, Dan Gusfield

5) Combinatorial Optimization: Algorithms and Complexity, by Christos Papadimitriou et al.