Sunday, February 10, 2008

Complexity Illness

Complexity Illness: One of the enduring stereotypes of academia is that people spend a great deal of intelligence, time, and effort finding complexity rather than simplicity. This is at least anecdotally true in my experience.[...]Who submitted a paper to ICML violating the 8 page minimum? Every author fears that the reviewers won’t take their work seriously unless the allowed length is fully used. (Via Machine Learning (Theory).)

These claims come from a highly idealized model of scientific research. Most good ideas start their public careers in a messy, overly complicated state. Scientific ideas, like computer programs, need to be refactored repeatedly by the relevant community before they reach the conciseness and clarity that John hopes for. In his wonderful collection Indiscrete Thoughts, Gian-Carlo Rota wrote:

Every mathematical theorem is eventually proved trivial. The mathematician's ideal of truth is triviality, and the community of mathematicians will not cease its beaver-like work on a newly discovered result until it has shown to everyone's satisfaction that all the difficulties in the early proofs were spurious, and only an analytic triviality is to be found at the end of the road.

My conference submissions often hit the limit because at the time of submission I haven't understood fully what I am proposing, or what experiments are the best to test what I am proposing. Ten years later, if the idea was good enough, it is likely that it and the most revealing experiments could have been written up in half the space. But the only reason that would be possible is that the work was published, and the debate that followed was crucial in distilling out what really mattered from the original rough brew.

As I tell anyone I am advising, the real value of publication is to contribute to the ongoing dialog of science, and learn from the responses more than we could ever learn by thinking alone in our offices. Dialog is messy, ideas come out half-formed, we hesitate, we track back, we speak over each other, we suffer from esprit d'escalier. Waiting to engage until we have the perfect formulation may reduce the chance of embarrassment, but it would also deaden a lively exchange, and take away delicious rejoinder opportunities for our debating opponents.

No comments: