radford-neal

5 ms·

1.
by radford-neal 1mo ago
Sorry, fumble fingering on my part...
2.
by radford-neal 1mo ago
.
3.
by radford-neal 2mo ago
Public opinion is not a binary. It could certainly move further against Trump, and just as significantly, it could move in the direction of free trade, whose opponents include plenty of people other than Trump. What's madness is contin
4.
by radford-neal 2mo ago
If a bear is rampaging around your campsite do you think you have to "take a stand"? This isn't some sort of status game. Well, maybe it is for Trump, but not for anyone who's not an idiot. Unfortunately, Carney is an id
5.
by radford-neal 2mo ago
Yeah. I'm not sure how one would handle that. (Can't remember if Whitehead & Russell had any inherent operator precedence.) So may be most useful if you don't have inherent precedence, as might be the case if you're
6.
by radford-neal 2mo ago
It's better visually. No dot, ., :, :., ::, ::., :::, etc. have increasing visual weight, which shows which are the more top-level operators without having to match up parentheses.
7.
by radford-neal 2mo ago
The notation for avoiding parentheses is interesting, and I've thought that it might be useful in programming languages. To illustrate, suppose you have a non-associative operator $. Rather than write a$(b$c), you can write a$.b$c - t
8.
by radford-neal 5mo ago
The river water picks up silt as it flows to the sea. The silt has to go somewhere. That's why the delta exists, and grows, absent other effects. The river wandering is part of this. Since the silt can only be deposited where the water
9.
by radford-neal 5mo ago
New Orleans is on a river delta, and without human effects on land, and without sea level rise, wouldn't one expect this delta to expand due to silt deposition? And would this silt deposition actually occur at a rate that would fully
10.
by radford-neal 5mo ago
Those interested in this might find my paper on "Representing numeric data in 32 bits while preserving 64-bit precision" to be of interest. Can be found at https://arxiv.org/abs/1504.02914 (note the code avai
11.
by radford-neal 6mo ago
The very first of the "false or unproven" claims they list is "animal protein is healthier than plant-based protein". And at this level of imprecision, that claim is true ! Animal protein has a good balance of amino aci
12.
by radford-neal 7mo ago
These are good points, but are mostly pragmatic, resulting from the rest of the system not being designed around a tree representation. Of course, that can be decisive in practice... A more fundamental issue (which I mention in the document
13.
by radford-neal 7mo ago
Well, I think most common transformations work reasonably well. One usually doesn't want to do things completely contrary to the AST, such as convert "while a<b" to "whilea = b". And it's certainly possible
14.
by radford-neal 7mo ago
Many years ago, I created an editor operating on syntax trees that I think is more "hard-core" than this - that is, only tree-oriented operations are done. There is no parsing of text, since entering plain text, rather than a tr
15.
by radford-neal 2y ago
I don't know what political machinations go on in the International Energy Agency, but one has to wonder why they thought it was useful to define a term "renewable" in this way, considering that such a term is of no use in fo
16.
by radford-neal 2y ago
Nuclear power is "generally accepted" as non-renewable amongst people opposed to nuclear power. Since uranium fuel is not going to run out on any reasonable time scale (and thorium will last even longer), it is "renewable&quo
17.
by radford-neal 2y ago
"If you want every CPU to support every instruction..," That does indeed seem nice. As long as there's an easy way to enquire about whether an instruction is slow or not, it means that all the software runs, and the software
18.
by radford-neal 2y ago
A recorder is a flute, more specifically, a fipple flute. But there are other fipple flutes, such as tin whistles. (One distinguishing feature of a recorder is the use of the thumb hole to control octaves.) So calling a recorder a &quo
19.
by radford-neal 2y ago
"Power-off data retention is increased from 1 month to 3 months." I don't think these are meant for ordinary users. One extended vacation and your data is toast...
20.
by radford-neal 2y ago
The court's ruling has actual legal effect. The inquiry's conclusion is, well, just somebody's opinion. One might also note that the invocation was never approved by parliament (it was terminated before the deadline for the
21.
by radford-neal 2y ago
A court has ruled that it did not: https://nationalpost.com/news/politics/a-judge-ruled-trudeau...
22.
by radford-neal 2y ago
A basic problem with evaluations like these is that the test is designed to discriminate between humans who would make good lawyers and humans who would not make good lawyers. The test is not necessarily any good at telling whether a non
23.
by radford-neal 2y ago
One hears stories of people with dementia having a day or two of mental clarity just before they die. Could this explain that? Perhaps their bodily processes are failing so much that normal glucose production is impaired, leading to keton
24.
by radford-neal 2y ago
"there is no guarantee that your preprint is accurate with the final, published version, so most people will still trust the paywalled version more than your PDF" I think this is backwards. The definitive version that should be c
25.
by radford-neal 2y ago
Since we're warm-blooded animals, maybe it's because IR vision would be swamped by the emissions of the eye itself?
26.
by radford-neal 2y ago
Exact addition using accumulators has progressed since the work of Zhu and Hayes in 2010. See my paper at https://arxiv.org/abs/1505.05571 (repository at https://gitlab.com/radfordneal/xsum ), whic
27.
by radford-neal 3y ago
The commenters here saying "humans cannot win anymore", "there is no chance that humans can beat the best Go AI anymore" are apparently unaware that this is no longer true . See https://arxiv.org/abs&#x2
28.
by radford-neal 3y ago
From 1992: https://core.ac.uk/download/pdf/44289908.pdf
29.
by radford-neal 3y ago
As the author admits at the end, this is rather misleading. In normal usage, "overfit" is by definition a bad thing (it wouldn't be "over" if it was good). And the argument given does nothing to show that Bayesian
30.
by radford-neal 3y ago
So why have a debt ceiling at all? A ceiling makes no sense if whenever it is about to be reached, unconditionally increasing it is regarded as imperative, with any attempt to link an increase in the debt ceiling to measures that would lim

More ›