Books | Jun-young Kwak

한국에서 가져가는 책들 :)

August 15th, 2009 § 0

좌안 – 마리이야기 1-2
우안 – 큐이야기 1-2
사랑후에 오는 것들 1-2
불안
The Winner Stands Alone
지금 사랑하지 않는 자, 모두 유죄

“Great scientists come in two varieties, which Isaiah Berlin, quoting the seventh-century-BC poet Archilochus, called foxes and hedgehogs. Foxes know many tricks, hedgehogs only one. Foxes are interested in everything, and move easily from one problem to another. Hedgehogs are interested only in a few problems which they consider fundamental, and stick with the same problems for years or decades. Most of the great discoveries are made by hedgehogs, most of the little discoveries by foxes. Science needs both hedgehogs and foxes for its healthy growth, hedgehogs to dig deep into the nature of things, foxes to explore the complicated details of our marvelous universe. Albert Einstein was a hedgehog; Richard Feynman was a fox.”

…From “Wise Man”, The New York Review of Books Volume 52, Number 16 · October 20, 2005

—————————————————————————————–

For any scientists, the question thrown by Freeman Dyson is not easy to answer. Even it is really hard to recognize their own type between foxes and hedgehogs.

This kind of question was always in my mind since my childhood, and frankly speaking, I wanted to be a hedgehog rather than a fox. However, I sometimes (well… frequently?) feel I am not fully ready to pursue my research like a hedgehog mainly due to lack of capability. Might being a compromise or consolation for me, I often used to think the life like a fox is also not too bad.

Exploiting vs Exploring.
Well… again, this is definitely one of hard problems, and even sometimes it is hard for us to differentiate them with a clear boundary. I do not think there is any good pure strategy to clearly tell one of them is superior than the other. Probably, the best way to resolve this issue is adopting a mixed strategy with some prior probability distribution (given by intuition?).

At this point, I think the best solution $(x^*, p^*)$ for me is:
$p^* \leftarrow argmax_p \{ r_{i} + \gamma \sum_{j=1}^N p_j^T \cdot x_j \}$
$x^* = p^{*T}\cdot \left( \begin{array}{c}Exploiting \\Exploring \end{array} \right)$
$where, p_{1} = \left( \begin{array}{c}0.2 \\0.8 \end{array} \right)$