srean

66 ms·

841.
by srean 8mo ago
You sir are keeping alive an old tradition. 1K BC old. https://en.wikipedia.org/wiki/Esagil-kin-apli#The_Sakikk%C5%...
842.
by srean 8mo ago
The standard way is to use libraries like libevent, libuv that wraps system calls such as epoll, kqueue etc. The other palatable way is to register consumer coroutines on a system provided event-loop. In C one does so with macro magic, or u
843.
by srean 8mo ago
> However, in other most languages, you have to read() in tiny chunks, parse, process, serialize and finally write() back to the disk. This works, but is verbose and needlessly limited C has those too and am glad that they do. This is wh
844.
by srean 8mo ago
Ah! Your comment helped me understand the parent comment so much more. I thought it was more about data hygiene needs. Yes a DT on raw pixel values, or a DT on raw time values will in general be quite terrible. That said the convolutional s
845.
by srean 8mo ago
First time hearing about it, so cannot comment without reading further. KGB and CIA had their noses in the affairs of every country, more so in the mid east, so cannot be ruled out. Historical interest aside, that was not the point I made o
846.
by srean 8mo ago
Oh theirs is a far more sophisticated and deeper idea, to look for invariant representations. Mine is a quick but effective practical hack fuelled by a little bit of insight.
847.
by srean 8mo ago
> It comes from first principles and historical truths. There is a book about it - it's the best selling book of all time. If you are talking about Euclid's Elements, it has certainly been a joy to work through them again in my
848.
by srean 8mo ago
I am talking about something that is from decades earlier. The Google search results I quoted summarises that quite well.
849.
by srean 8mo ago
As long as divide and rule is employed it won't.
850.
by srean 8mo ago
To a certain degree yes. https://en.wikipedia.org/wiki/Esagil-kin-apli In this Mesopotamian text, diagnostic rules are structured as a nest of if then else rules. So I have been told, not that I have read it myself.
851.
by srean 8mo ago
Ah yes that was it. Thank you.
852.
by srean 8mo ago
That has not been my experience though, apart from the need of standard data hygiene that one has to maintain for any ML exercise. Normalising the numeric features to a common range has been adequate. This too is strictly not necessary for
853.
by srean 8mo ago
Which successful country does not covertly monitor and try to influence the politics of neighboring non-ally countries ?
854.
by srean 8mo ago
To the contrary the Ayatollah has (now had) a formal and public fatwa against nuclear weapons.
855.
by srean 8mo ago
What I am claiming isn't even controversial. It's quite well known and accepted. So much so all the standard search engines will confirm it, and search engines typically reflect a consensus view of crawlable documents. This is wha
856.
by srean 8mo ago
Yes you can certainly do that. What I had not articulated well is that linear classifiers have the opportunity to use matvecs that have a different level of L1 L2 cacheable goodness and non-branchy code. There using proper memory layout giv
857.
by srean 8mo ago
Well that would implicate the state of Israel as well. It's Israel that set up Hamas in the 1970s to split the support of the largely secular PLO.
858.
by srean 8mo ago
Now that you ask, the name eludes me. Am I misremembering ? was he an MP who lost office? I certainly remember a flurry of news reports.
859.
by srean 8mo ago
I have been quite impressed by UK politics. Entertain paedophiles ... land up in jail. Be rude and patronizing to women, there goes your prime-ministership. Keep it that way please. Accountability feels so rare in recent US politics.
860.
by srean 8mo ago
Taken to the limit you are absolutely right. It turns out many dataset have such a fractal like nature but where the partitioning needs to be cut off at a certain depth and not continued till infinity.
861.
by srean 8mo ago
Amen. To support your point of view, I haven't encountered this in particle physics. It's from other branches. I am not a Physicist myself, happened to be in a position to observe funding requests, request for collaborations.
862.
by srean 8mo ago
Yes. I mention restricted oblique trees in passing in my original comment. In my experience, oblique trees tend to add considerable complexity, the others more so. Of course whether the complexity is warranted or not will depend on the data
863.
by srean 8mo ago
In general I would agree. I think it holds true at the highest levels. What worries me is the noticeable uptick of presentations of the sort -- look ma better fit ... deep neural nets. These are mostly by more junior folks, but not necessar
864.
by srean 8mo ago
I missed a word "recursively", that I have edited in my original comment now. Consider connected regions in the domain that have the same label. Much like countries on a political map. The situation where this has a short descript
865.
by srean 8mo ago
> Times have changed… This makes me a little concerned -- the use of parameters rich opaque models in Physics. Ptolemaic system achieved a far better fit of planetary motion (over the Copernican system) because his was a universal approx
866.
by srean 8mo ago
For linear algebraic transformation applied to several rows at once, I wholeheartedly agree. Not so convinced about decision trees though (that process one row at a time). Yeah, unless you had to deal with arbitrarily large integer features
867.
by srean 8mo ago
A 'secret weapon' that has served me very well for learning classifiers is to first learn a good linear classifier. I am almost hesitant to give this away (kidding). Use the non-thresholded version of that linear classifier output
868.
by srean 8mo ago
What benefit does numpy or dataframes bring to decision tree logic over what is available in Guile already ? Honest question. Guile like languages are very well suited for decision trees, because manipulating and operating on trees is it&#x
869.
by srean 8mo ago
You have this thing a little backwards that it is unintentionally hilarious. Decision trees predate KD trees by a decade. Both use recursive partitioning of function domain a fundamental and an old idea.
870.
by srean 8mo ago
You can search for the "peeking" problem in A/B testing. SPRT also very likely helped win a major war that involved many nations.

More ›