LolWolf

47 ms·

451.
The Desperate Battle to Destroy ISIS(newyorker.com)
9 points by LolWolf 10y ago|0 comments
452.
by LolWolf 10y ago
Sure, you can add an oracle for axiom of choice and you're done (note that AoC -> Excluded Middle). A proof is then "constructive" in the same way (e.g. Can be also written as a program, there).
453.
Metric mishap caused loss of NASA orbiter (1999)(edition.cnn.com)
1 points by LolWolf 10y ago|0 comments
454.
by LolWolf 10y ago
Thank god I'm not the only one who did this on a first read.
455.
by LolWolf 10y ago
I disagree. Why can we not just get really really good at measuring the universe and describe everything from there? If you agree with physics as a decent approximation of the universe, everything else is collective/emergent (though
456.
by LolWolf 10y ago
Certainly will; thanks again for the interesting post!
457.
by LolWolf 10y ago
I disagree with the above, but I think I can shed light on what they might mean. Usually, control theory (which is used in most manufacturing processes) requires quite a bit of background knowledge on the processes at hand along with fairly
458.
by LolWolf 10y ago
The whole deal seems weird to me. Neural networks are just functional approximators, so why isn't a linear regressor of k-th order (e.g. Taylor expansion up to k-th order) also considered "ML"? What's the distinction her
459.
by LolWolf 10y ago
Or his company Lattice for that matter.
460.
Mark Zuckerberg and the Internet(medium.freecodecamp.com)
1 points by LolWolf 10y ago|0 comments
461.
by LolWolf 10y ago
Sure, I agree; in this case I refer to "tools" as theoretical tools rather than computational ones. There's no debate on the merits of computational tractability by converting problems to linear algebra ones; I was speaking f
462.
by LolWolf 10y ago
These are all well-known things in simplicial geometry [1] which is used everywhere in computational topology and discrete differential geometry. It turns out all of these operations can be done efficiently by considering a matrix over F_2
463.
by LolWolf 10y ago
I agree: I don't really see what the obsession is with linear-algebra based graph theory. Yes, it's cool, yes it has interesting results, but it's not everything and we've got tools for semigroups and general algebras th
464.
by LolWolf 10y ago
I should be more careful: this operation does, in fact, form a lattice if the graph is connected at every point; e.g. the + operation is restricted.
465.
by LolWolf 10y ago
No, Lin-Alg is in fact what I work on in my research on spectral graph theory and graph partitions for approximation algorithms; but I do think there are different tools for different jobs. There are plenty of interesting results where line
466.
by LolWolf 10y ago
Prove that a topological sort exists for a given graph using only linear algebraic properties of the adjacency matrix or the laplacian. Or even, give a proof that a graph is a DAG iff there exists a topological sort using only linear algebr
467.
by LolWolf 10y ago
I do agree with the anti-snobbery part, but, again, as someone who totally skipped over the Haskell parts (which were mostly gibberish to me), I still think the article was quite enlightening. Of course, it's also a matter of the audie
468.
by LolWolf 10y ago
I'm afraid I don't quite understand what the negation of a graph is in this case and what you mean by 'overlay,' perhaps elaborate a bit?
469.
by LolWolf 10y ago
As far as I can tell, I would have to answer that in the negative (see the EDIT in the previous post). In general, the set of graphs is more complicated/general than the set of countable/finite lattices (every lattice can be repre
470.
by LolWolf 10y ago
Yep, I definitely agree! Nominally, there are plenty of interesting results in spectral graph theory which make use of operations like these (esp. when computing laplacians/spectral partitions, etc), but, like most tools, while we can
471.
by LolWolf 10y ago
That's fine, but then in the 'infinite vertices case' we require an associated numbering scheme and defining an ordering over vertices which requires a bijection from whatever set you're working with into an index labeli
472.
by LolWolf 10y ago
Great, thanks!
473.
by LolWolf 10y ago
You're actually very much right, in this case, if we choose z to be the preferential element (say, the largest, such that x->y iff x≤y), the addition operation would be the same as 'adding' elements/orderings to the p
474.
by LolWolf 10y ago
In particular, you can also endow graphs with lattice structures, which yields a bunch of of nice monotonicity/fixed-point/ordering theorems over graphs; computability/hardness of these properties touches some parts of my res
475.
by LolWolf 10y ago
Matrices under addition? I know very little of Haskell, but I think this blog post is quite interesting; it's not obvious to me how you would translate the 'join' operator (->) into a matrix formulation that preserves nice
476.
by LolWolf 10y ago
This article has been reposted a few times [1][2]. Is there any way to merge discussions? ------ [1] https://news.ycombinator.com/item?id=12317525 [2] https://news.ycombinator.com/item?id=7928924
477.
by LolWolf 10y ago
It depends on what kind of analysis you're doing. If you're doing pure exploratory stuff on large datasets, you'll mostly be using NumPy as a tool to help with Pandas. See McKinney's Python for Data Analysis (I believe
478.
by LolWolf 10y ago
I'm a huge fan of Python and I use it for everything (to put it mildly, I've spent weeks translating MATLAB code from other researchers into Python simply because I believe Python is far nicer to work with), but I must say that
479.
by LolWolf 10y ago
Go buy it on Steam!
480.
Unexpected hanging paradox(en.wikipedia.org)
2 points by LolWolf 10y ago|0 comments

More ›