I think that what MIT considers innovative and what Harvard Business schools thinks is innovative are different, and the models for getting there are going to be different from each school of thought. The HBR article touches on some things that produce what I would consider innovation: Researchers and scientists using the scientific method, but it never goes that far.

There are discussions here about Xerox PARC, and I would suggest including Bell Labs and probably a few other business research parks, and you'll find highly educated people given the freedom to experiment.

I guess my argument is that in order to produce innovation, I think that the best thing that management can do is put the right people together and get out of the way, but undirected innovation might not aid the business. It's an interesting problem.

EDIT: I should include the processes that developed things like Linux.