Youjun Hu

computational plasma physicist,
machine learning practitioners,
enjoy learning, writing, programming.

[Youjun Hu @YoujunHu]

Animation of n=6 TAE saturation and induced energetic particle heat flux https://t.co/Y0kkxBQvGF via @YouTube

[Youjun Hu @YoujunHu]

python's usage of negative index makes addressing negative frequency of FFT much easier.

[Youjun Hu @YoujunHu]

I can only write freely in the appendix of my paper.
In the main text, I feel that I need to be brief, and need to skip details.
So I feel urged to move the details to the appendix.

I turns out the details are what I love.
So I enjoy writing appendix, rather than the main text.

[Youjun Hu @YoujunHu]

@paulg I just read all the footnotes together after I read an essay by searching, e.g., [1, to know the context of the footnote.

[Youjun Hu @YoujunHu]

Am I a pusher? https://t.co/U9rpzKax3n via @YouTube

[Youjun Hu @YoujunHu]

This is my youtube chanel: https://t.co/XDwYfNQouA via @YouTube
I plan to create more video in futures.

[Youjun Hu @YoujunHu]

but for noisy signals, there are many local peaks, which would make this method inaccurate.

[Youjun Hu @YoujunHu]

when I need to calculate growth rate of an oscillating signal, I often do it manually (by taking logarithm and fitting a line to it). I turns out python can do it so nicely:

[Youjun Hu @YoujunHu]

I'm biased, and that's good. Because I need a working hypothesis (which may be wrong) to begin with, then I can iterate and improve.

[Youjun Hu @YoujunHu]

simplifying in order to understand, to find bugs, not for the purpose to be practically useful. For practical usage, we need complex system consisting many subsystems, taking into many factors, not just some dominant factaor.

[Youjun Hu @YoujunHu]

"We learn more from failure than success."
"We learn more from writing than reading."
"We learn more from building than watching."

[Youjun Hu @YoujunHu]

These physics can be captured in the measured constraints. So reduced model can be more useful in practice. This is the main reason I should do integrated modeling.

[Youjun Hu @YoujunHu]

the first-principle model is not necessarily more accurate than reduced 0D or 1D model because the reduced models can incoporate measured constraints easily in the model, but the first-principle model can miss some important physics that we do not know we are missing.

[Youjun Hu @YoujunHu]

one method to measure your contribution to the world is to check whether google AI cites your works as a reference in its AI generated content.

[Youjun Hu @YoujunHu]

I disliked writing untill I found I can write for myself.

Always write for myself to read, to understand, to grow from there.

to remind myself.
to clary something for myself.
to record somthing for future me in case I forget them.

[Youjun Hu @YoujunHu]

Meaningful work can makes people happier.
That's one of the reasons that I should do integrated modeling, which seems meaningful to me at present.
Is gyrokinetic simulation also meaningful?