imurray

12 ms·

91.
by imurray 8y ago
Dense matrix algebra has the benefit that besides the matrix dimensions, the data itself has no impact on performance I was very surprised when I first learned that's not strictly true, because of denormal numbers[1]. Here's a s
92.
by imurray 9y ago
"Verlag" is just the German word for publishing house.
93.
by imurray 9y ago
Why link to the pdf? The webpage https://web.stanford.edu/~jurafsky/slp3/ links directly to the PDF, and gives context and other download options. It's not so easy to go back from the PDF to the webpage. Mods
94.
by imurray 9y ago
You might want to search with / -p
95.
by imurray 9y ago
I suggest changing the link from the .pdf to the web page: https://papers.nips.cc/paper/7181-attention-is-all-you-need It's one click to get the pdf from there. But you also get a plain webpage with abstract, cita
96.
by imurray 9y ago
if you do inline quoting, MS Outhouse web client thinks that everything after the beginning of the quotation is a copy of the previous message and folds the whole thing, meaning your recipients get back what looks like an empty message. Re
97.
by imurray 9y ago
Coincidentally I just went to a nice exhibit in Edinburgh co-authored by Henry Segerman: http://www.maths.ed.ac.uk/~aar/brilliantgeometry/ — the last day is tomorrow for Edinburghers. He has a book with supporting
98.
by imurray 10y ago
Implementations of HMC are approximating ideal Hamiltonian dynamics by discretizing time. The "symplectic" integrators that are used have the property that they correspond to simulating continuous-time Hamiltonian dynamics exactl
99.
by imurray 10y ago
Yes, you need the gradient. And despite what the autodiff people say, that might be hard and/or not very useful in practice, if your likelihood is based on running a lengthy astrophysical simulation. You can run HMC on any differenti
100.
by imurray 10y ago
I think you're assuming that "automatic differentiation" (AD) is the same thing as "using dual numbers", but AD is a family of methods including dual numbers. The blog post you link to says they're using https
101.
by imurray 10y ago
> If I had to guess, I’d say that dual numbers may be slightly slower than backpropagation The cost of Dual numbers (a form of forward-mode differentiation) scales linearly with the number of derivatives (just like finite differencing,
102.
by imurray 10y ago
If the function is analytic and supports complex numbers (usual in Matlab/Numpy/Fortran/...), but you're only interested in the real part, a quick hack is to abuse complex numbers and compute imag(f(x+i*epsilon))/ep
103.
by imurray 10y ago
The article nicely demonstrates that .jpg is not a good format for screenshots of text.
104.
by imurray 10y ago
Neural translation has been really amazing and impressive. I stumbled into a way to be mean to Google translate this week: https://twitter.com/driainmurray/status/818934530207862784 OpenNMT can also run amok given
105.
by imurray 10y ago
https://news.ycombinator.com/item?id=11862476
106.
by imurray 10y ago
You could possibly try https://twitter.com/AlisonBLowndes
107.
by imurray 10y ago
Would delete this post if I could. Was a request to fix a broken link. Now fixed.
108.
by imurray 10y ago
I'm going to check back later to see if anyone manages to reproduce it. Perhaps by the time it's presented at NIPS. A twitter conversation reflecting some scepticism, but agreeing it would be interesting if it all checks out: htt
109.
by imurray 10y ago
Right-clicking on the image and selecting "View Image" (Firefox), or "Open image in a new tab" (Chromium), gives the webpage, not the image. I can see why that happens: the menu items just open a URL and don't force
110.
by imurray 10y ago
I agree: If you have data X and outputs y and must have a linear function of X to minimize square error on y and only y, then yes we have defined our task. I think it's a lucky engineer that has their end-goal defined as such a crisp m
111.
by imurray 10y ago
I'm not sure what you mean, but I think we're talking past each other. Regression is a statistical term. If it's going to be used for an application involving data and real decisions, not understanding the statistical issues
112.
by imurray 10y ago
If you're new to learning about regression and want to use it, I would prioritize the statistical issues. There are lots of computational details you could read about, like in this post. However in simple cases, fitting least-squares l
113.
by imurray 10y ago
The thinning here is based on a misconception. If the purpose is plotting a histogram it doesn't matter if the samples are correlated. The bin heights are consistently estimated if you keep everything. Throwing away intermediate steps
114.
by imurray 10y ago
A lot of machine-learning papers are eight pages. Speech conference papers (heavy users of neural nets) are often only four. Some details aren't part of the main message, so don't make it in. Often code is available, and initializ
115.
by imurray 10y ago
Nice. Recreating these methods in simple code for yourself is definitely the way to check you understand it. This demo looks nice, clean, and straightforward. (Although I'd rename or comment variables x and y, or give some sort of guid
116.
by imurray 10y ago
No it's not the same amount of clicking. From Gimp 2.7 to 2.8 something that used to be ctrl-s and close window went to ctrl-shift-e, and answering four dialog boxes before getting the image saved and the window closed. Around the time
117.
by imurray 10y ago
I totally see that it makes sense for some work flows, including the ones Gimp is probably most meant for. It was frustrating that for the people whose workflows were messed up, the developers were totally opposed to doing anything about it
118.
by imurray 10y ago
After ctrl-shift-e I have to hit enter in response to different pop-up dialogs three times, and then still answer a fourth confirmation when I quit gimp. If I'm editing a bunch of files that's not fun. With the Saver plugin I ha
119.
by imurray 10y ago
Gimp 2.8 frustrated me: I could no longer open a .png, edit it and easily resave it back to .png. The Gimp developers knew better and made the UI strongly favor saving as .xcf, which makes sense in some use-cases, but not mine. I found this
120.
by imurray 10y ago
Don't forget the biggest edge Python has over MATLAB for numeric simulation, i.e., you can prototype your work as quickly, and for the same program, convert the slow portions into C or Fortran without the need to rewrite the whole thi

More ›