LolWolf

40 ms·

361.
by LolWolf 9y ago
I'm... not really sure what argument is being presented here. There's some weird accusations of vagueness, yet I see no proposition for a definition of free will here (and in the original comment, there was never attempt to provid
362.
by LolWolf 9y ago
Not that I agree with the GP, but non-determinism doesn’t imply free will in the usual strong sense of the phrase (e.g. that there is some intrinsic notion that humans can directly dictate their own actions). Non-determinism (as in QM non-d
363.
Modulating fluid intelligence through cognitive training and brain stimulation(sciencedirect.com)
2 points by LolWolf 9y ago|0 comments
364.
The strategy-proofness of FIFA and UEFA tournament qualifiers(arxiv.org)
1 points by LolWolf 9y ago|0 comments
365.
by LolWolf 9y ago
One can only make keys so large to be useful—my point is the same, still, that one would only have to break/find/buy/steal a single key instead of millions.
366.
Simple random search provides a competitive approach to reinforcement learning(arxiv.org)
1 points by LolWolf 9y ago|0 comments
367.
by LolWolf 9y ago
Or anyone who is willing to spend enough computational power now only needs to crack one key instead of millions of different ones.
368.
by LolWolf 9y ago
I was recently and undergrad (now a grad student)---I've done research, work, taught classes, led competition teams, etc... and I still struggle with this. You should definitely write a book!
369.
by LolWolf 9y ago
Also, sorry, I was taking terminology from the GP. The real term for this is cross-correlation —with autocorrelation being the application of the above operation onto a single function; i.e., taking ac(f(t), g(t)) as defined above, the a
370.
Shortest path search on time-dependent graphs(guille.site)
1 points by LolWolf 9y ago|0 comments
371.
by LolWolf 9y ago
The symmetry is mostly an effect of 'natural' images not favoring a specific axis (i.e., vertical vs horizontal lines are still distinguished as the same 'type' of lines, in some sense). Another way of seeing why the sym
372.
by LolWolf 9y ago
It's a pretty minor difference, really. Over functions (which is the same as over vectors/matrices, except the latter are taken at discrete points), the autocorrelation and the convolution differ by a sign. The definition of convo
373.
by LolWolf 9y ago
There are... one of them is given in the article—which calls the current system "one-shot"—there exists k-shot variants which allow for at least k opportunities/warnings before automagically braking. The (apparent) reasoning
374.
by LolWolf 9y ago
Maybe; I mean, see [0] where an accidental experiment was made when a container with >28k rubber ducks and other floating toys, tipped. New evidence for some models was discovered using roughly the same idea—by tracking the floting toys
375.
by LolWolf 9y ago
Oh, interesting, did not know that! I guess I only now just realized that 24d10 is 10 dice with 24 values (I just read the GGP and thought, oh 240 uniform is easy to make)... there's probably some nice way to emulate that with 1-bit
376.
by LolWolf 9y ago
Well, you can re-flip on >240, with an expected number of flips to receive a number in the range being around 1/(1-15/256) ~ 1.062, so pretty close to just one 8-coin-flip (with some low probability of having to re-flip, which
377.
by LolWolf 9y ago
I mean, ideally I think it would be written as an FPGA-like piece of hardware where the connections can be synthesized from a high level description language. A la [0]. —— [0] https://web.stanford.edu/group/brainsinsili
378.
NN for Predicting Results of Asking Barry Cottonfield to Prom in 1997 (2017)(arxiv.org)
1 points by LolWolf 9y ago|0 comments
379.
by LolWolf 9y ago
Probably to not be on record of the ATC, would be my guess.
380.
by LolWolf 9y ago
Note: this is just a general guess. But my money is on this just being a military test—some SR71-like replacement (probably air-breathing, maybe even unmanned). It's not really out of the question, but a big hole in this idea is that m
381.
Pigovian tax(en.wikipedia.org)
2 points by LolWolf 9y ago|0 comments
382.
by LolWolf 9y ago
Then you can only cross at extrema.
383.
Comparing classical optimization vs. reinforcement learning for autonomous drones(guille.site)
1 points by LolWolf 9y ago|0 comments
384.
by LolWolf 9y ago
Stanford continues to own the land—I think they almost exclusively lease it (even if it is long-term).
385.
Solving SVMs using proximal gradient methods(guille.site)
3 points by LolWolf 9y ago|0 comments
386.
The South Sea Company(en.wikipedia.org)
2 points by LolWolf 9y ago|0 comments
387.
Tulip mania(en.wikipedia.org)
1 points by LolWolf 9y ago|0 comments
388.
by LolWolf 9y ago
This approximation has vanishing error from a linear function (as the interval decreases) and (most importantly!) we only have finite noisy samples; so it’s essentially indistinguishable (in a statistical way) for a small enough interval gi
389.
by LolWolf 9y ago
This is pretty fascinating! Though the concept seems to work only for convex problems (in particular, problems which have strong duality; this excludes NNs in almost their entirety, except 1 layer nets), but the application is nice and stra
390.
by LolWolf 9y ago
Just for funsies (but may answer your question): every (differentiable) function is linear at a small enough time scale. This includes e^x or pretty much every function you listed.

More ›