How do I become a data scientist?. [link]

It takes a few courses and reading a few posts to become one. Great replies from both academia and industry's perspective.

Machine Learning that Matters. [Paper]

An unusual but inspiring invited talk at ICML'12. It makes a lot of sense to always connect research to the needs of the real world. [slides] A short summrizing comment of the talk can be found at this post.

Hollow face illusion: a powerful illustration of motion cue in human's 3D vision. Got a piece of this on my desk to remind me what we see is not necessarily what there are. "Seeing is NOT believing".

Richard Hamming: You and your research. [link]

What does it mean to do great research? What can be learned from Shannon, Hamming and fellow scientists at the golden age of Bell lab? An insightful speech, rather different from those cliche theory of success.

SCIgen - An automatic paper generator from MIT [link]

An open source software to generate plausible-looking CS papers based on the structure and vocabulary of a typical CS paper. To date, SCIgen has successfully generated quite a number of "random" papers that get accepted by conferences and journals. Just for fun.