jmmcd

37 ms·

361.
by jmmcd 10y ago
Fascinating! The methods are very interesting -- a threshold (similar to "just noticeable difference" (JND) used in cognitive/perceptual science); an attempt at estimating the size of the space; an attempt using Hausdorff d
362.
by jmmcd 10y ago
> feel the vibe! the best shocks [...] Users enjoying electric shocks. Reality distortion field operating at nominal.
363.
by jmmcd 10y ago
Not working in Ireland.
364.
by jmmcd 10y ago
Don't forget that Douglas Adams always was a huge Apple fan. Bearing this in mind, I read comments like that one as a dig at non-Apple computers of the time.
365.
by jmmcd 10y ago
> Cook: [...] Apple has long supported international tax reform with the objectives of simplicity and clarity. > Gardner: That bit made me laugh out loud. When the Senate’s permanent subcommittee was describing the elaborate tax-avoid
366.
by jmmcd 10y ago
Lasagne on top of Theano https://github.com/Lasagne/Lasagne
367.
by jmmcd 10y ago
Hmm.. the entry on inflation hasn't changed (doesn't say anything like what I said) but I see notes that mention grade inflation. I don't know whether that was there first time round.
368.
by jmmcd 10y ago
I think these can be defended as mental models, even though the article doesn't do a good job of it. We're familiar with inflation in the financial sense. But then there is also grade inflation. There is inflation of superlatives
369.
by jmmcd 10y ago
Part of the problem in anomaly detection is that sometimes anomalies are of a "novel" type. It can be impossible to collect labelled data on all possible anomalies. So learning a model of normal behaviour turns out better. But if
370.
by jmmcd 11y ago
If the printing press or web were proposed as research topics by a new PhD student today, they'd probably be laughed out of the room. But if they were proposed by someone with credibility, who knows what research is like from top to bo
371.
by jmmcd 11y ago
No, I think pure research can be and usually is ordinary research. I agree with your points about the young and non-established being good bets for big thinking. But I think the PhD is an unsuitable vehicle, partly because it's not wel
372.
by jmmcd 11y ago
Diverting resources sounds good, but obviously should come first. The research itself could and should have value. But a PhD must go all the way to the bottom, in terms of detail. It's not a pop science book or a funding programme, w
373.
by jmmcd 11y ago
Great! Meanwhile his PhD was not particularly "big thinking". So in Einstein's case I think we're both happy? As for jobs jobs , I think you missed the point. If we lie to PhD candidates about career prospects and train
374.
by jmmcd 11y ago
> We need big thinkers We need some, but there aren't anywhere near enough big-thinking jobs for all these PhDs. > The success of science and technology will always lead to new ethical dilemmas. How do you spot them and try to wo
375.
by jmmcd 11y ago
> It sounds to me like this isn't a great match for what you're looking for. Do you agree? This is the part that reveals the whole thing as nonsense. My "ideal job" is not what I'm "looking for", theref
376.
by jmmcd 11y ago
Is there a version including the figures?
377.
by jmmcd 11y ago
Lovely! It reminds me of some methods of generating images with genetic programming, e.g. Karl Sims, http://dahart.com , and http://picbreeder.org . The main difference being a stack-based representation versus a tree
378.
by jmmcd 11y ago
The challenge was recently discussed on HN: https://news.ycombinator.com/item?id=9874215 This new article is from the lecturer's point of view.
379.
by jmmcd 11y ago
The players don't have to be rational. It is interesting to study PD with non-rational players, e.g. when it arose as a model of the nuclear arms race. It's true that the PD also "shows why two purely 'rational' i
380.
by jmmcd 11y ago
Everyone can work that out for themselves! In theory. Or else remember it from the course textbook. In context, "coordination" usually means communication after the game is set up, whereas the idea of the random number generator i
381.
by jmmcd 11y ago
I'm afraid you need to re-read your notes :) N-player prisoner's dilemma is well-known. Neither 2-player nor N-player assumes rational actors. EDIT: It assumes that payoffs are in a reasonable unit of utility (so we don't nee
382.
by jmmcd 11y ago
SPOILER? The "correct" answer is to generate a random number in [0, 1]. If less than 0.1, then select 6, else select 2. For safety use 0.05 (say) instead of 0.1. This means that everyone is guaranteed 2, but some (as many as possi
383.
by jmmcd 11y ago
Lots of libraries won't work with Brython either.
384.
by jmmcd 11y ago
One thing is for sure, changing alpha is not the answer! I thought the article was going to be about six sigma and Japanese theory of manufacturing. For a while the article was going there, with a mention of changing incentives and rewardin
385.
by jmmcd 12y ago
The fourth image on the DK effect is identical to the third. Copy-pasted link, maybe.
386.
by jmmcd 12y ago
Yes, but it still makes no sense to compare GA performance with a gradient-based method on the problem GP is talking about, because an explicit gradient is available.
387.
by jmmcd 12y ago
It's not gradient based because it runs without reference to the derivative (gradient) of f. It's black box. Gradient descent does not work by generating random neighbours, it follows the gradient.
388.
by jmmcd 12y ago
If gradient is available then it makes little sense to say you want people to compare that method to GAs. GAs are for black box search.
389.
by jmmcd 12y ago
"Shotgun" just means "with restarts" (right?) but what do you mean by "ballistic"? "With momentum" maybe? I can't see that term used anywhere in the literature.
390.
by jmmcd 12y ago
> In general, p-values are based on what would have happened under other possible data sets. As a hypothetical example, suppose a researcher is interested in how Democrats and Republicans perform differently in a short mathematics test w

More ›