nonbel

56 ms·

721.
by nonbel 8y ago
Basically when talking about multiple particles, and using the term source, it would be clearer to say "source of x". Yet three times now you've just said "the source". It seems needlessly confusing so I think there
722.
by nonbel 8y ago
> "But it’s worth noting that for the most part, people already had tools for analyzing data. They came in the form of SAS, Stata, SPSS, Minitab, Microsoft Excel, and my personal favorite, XLisp-Stat (thanks Luke Tierney!). But the
723.
by nonbel 8y ago
>"The protons are accelerated at the source." So by "source" you mean "source of the protons"? And "local photons" means "produced near the source of the protons"?
724.
by nonbel 8y ago
So if I know the source but you don't, its local to me but non-local to you?
725.
by nonbel 8y ago
Not saying you are wrong about the terminology, but if local photons are "produced at the source", the use of "local" seems to be redundant. When would a photon not be created at the source? What is a non-local photon, o
726.
by nonbel 8y ago
>"After 3.9 billion years of hurtling unhindered through the vast reaches of the universe, a ghostly neutrino particle died on 22 September 2017. It was annihilated when it collided with an atom in the frozen darkness two kilometers
727.
by nonbel 8y ago
While outright fraud is probably rare, "questionable research methods/practices" are standard. Most people doing them dont even know there is something wrong with it. Pick any biomed or social science paper and I will quickly
728.
by nonbel 8y ago
Do you really consider this an obstacle? Obviously you need to have a validation set for the optimization and a new/unused test set for estimating out of sample performance.
729.
by nonbel 8y ago
Sorry, its still unclear. Do you consider genetic algorithms, etc to be an example of (mixed-integer) non-linear optimization? Or are you talking about some other method? >"You can probably do this kind of hyper^n optimization as de
730.
by nonbel 8y ago
Didnt you make the exact same post here? https://news.ycombinator.com/item?id=17497235
731.
by nonbel 8y ago
What are you saying goes beyond "101 optimization algorithms"? Are you saying to put another "layer" of deep learning to tune the parameters of the first "layer"? Doesn't this just add even more parameters
732.
by nonbel 8y ago
This is everywhere in academia. Peer review is like the goggles in the face of a seemingly neverending wave of BS: https://www.youtube.com/watch?v=juFZh92MUOY Honestly who cares about these papers in the case of ML though?
733.
by nonbel 8y ago
Yea the site looks a bit disorganized and crackpotty. Perhaps because its translated from french? You can find videos of them on youtube actually making these stone blocks though. The method works. Based on how long it took a small team to
734.
by nonbel 8y ago
Supposedly there is, "Irtysen Stele", "Sehel Stele" and "Famine Stele": https://www.geopolymer.org/archaeology/pyramids/pyramids-2-t...
735.
by nonbel 8y ago
You don't have to move huge stones. Its possible to cast limestone in place, like concrete. I read/watched some stuff about this guy not too long ago: https://en.wikipedia.org/wiki/Joseph_Davidovits He claims
736.
by nonbel 8y ago
>"Thinking about what kind of science one might do around evolution, the neutral theory seems like a very natural thing to test against. It seems useful to disprove that a hypothesis might be part of a separate or competing theory.&
737.
by nonbel 8y ago
Great to hear! Here is another you may find interesting: Meehl, Paul E (1967). Theory-testing in psychology and physics: A methodological paradox. Philosophy of Science, 34(2), 103–115. http://www.jstor.org/stable/10.23
738.
by nonbel 8y ago
>"Your analysis tells nothing about whether the model of fair coin being flipped is confirmed. You see a deviation of 0.06 in the mean and 0.47 in the variance. Is that negligible?" Th easiest way is to just run a simulation th
739.
by nonbel 8y ago
Why not? I mean I could go find examples of Ronald Fisher (guy who popularized them) doing this if you want...
740.
by nonbel 8y ago
I'm not saying get rid of the null hypothesis. I'm saying set the "interesting hypothesis" as the hypothesis to be "nullified" (ie, tested).
741.
by nonbel 8y ago
>"When model predictions are statistical, then, you want to answer the question: how "likely" is this model, given the data?" Huh? Here is a model with statistical predictions. Its a model of flipping a fair coin 20 t
742.
by nonbel 8y ago
>"Furthermore, the neutral theory is not merely a neutral model, to be used as a null hypothesis against which more interesting hypotheses can be tested." What possible scientific purpose would this serve? Why not just test the
743.
by nonbel 8y ago
Why are there so many people posting in this thread who seemingly want this project to fail? What do you have invested into it failing? To me this seems like a cool experiment, why not try it out?
744.
by nonbel 8y ago
I don't know of any examples. I was hoping you did.
745.
by nonbel 8y ago
>"I surely may be wrong in my prediction, but the thing I'm most fascinated about by these libertarian settlement plans is when they organically reintroduce all the government systems they wanted to escape in the first place.&q
746.
by nonbel 8y ago
I just disagree with the claim that understanding the mathematical machinery is necessary for tuning hyperparameters. I doubt it is even helpful.
747.
by nonbel 8y ago
Where did I say not to try to understand the math behind it? I say its not required to tune hyperparameters. To me, its doubtful that someone whose first impulse is to generate strawman arguments is doing a good job in analyzing data. Thats
748.
by nonbel 8y ago
For example look at this tutorial about a regularization hyperparameter: https://medium.com/data-design/xgboost-hi-im-gamma-what-can-... Id think this is much more useful than anything about the math. How much can you
749.
by nonbel 8y ago
Besides having a basic understanding of what the parameters do (this is depth, this is learning rate, etc) I don't see what insights are to be gained. The optimal parameters depend on the particulars of your dataset, that's why ev
750.
by nonbel 8y ago
> "The problem is that understanding all of the mathematical machinery is tricky and, unfortunately, these details are needed to tune the hyper-parameters." You don't need to understand anything about the math to run a ra

More ›