The (rather inevitable) ending, that is, where nice == brutal
3368 posts • joined 24 Apr 2007
The (rather inevitable) ending, that is, where nice == brutal
The librarian concurs
This mission has brought so many awesome images, and many, many other results. I remember being spellbound by the footage of the descent of the Huygens probe on Titan. Big thumbs up to all involved in making this possible. I will be raising a glass in memory of a great mission when Cassini has performed his death dive
I wouldn't recommend boiling oceans, as it leaves the fish rather dry and tasteless. Poaching is always better, especially for white fish
Sorry, couldn't resist. Mine's the one with the Keith Floyd cookbook in the pocket
Somehow, Tom Lehrer's "Masochism Tango" springs to mind. In particular the snippet:
"All at your command,
before you here I stand,
my heart is in my hand ...
... or perhaps mrs Alice Beeblebrox, 108 Astral Crescent, Zoovroozlechester, Betelgueze V
Vodka is a good addition, but I would also like some bacon
(just, because, well, ..., it's bacon!!)
"I suspect their eyes are permanently damaged unfortunately. "
It depends a bit on the duration and level of exposure. Most people return to normal vision within 1 - 6 months, fortunately. In many cases the damage to the retina appears to be reversible (more-or-less like a sunburn to the retina). Only in fairly rare cases is the damage permanent. As you correctly state, in those cases the brain can help "work around" the damaged bits, but if the damage to the foveal area is severe in both eyes, permanent loss of visual acuity may well result.
Let's hope they fall into the larger, first category
I cannot verify if the plaintiffs' eyes are really damaged, but it is certainly possible (hopefully it is only temporary damage, no worse). This is why I (and all others I know doing solar astronomy outreach) always warn people to buy directly from a reputable astronomy store (as I did for the family). Only during totality is it safe to watch the sun without protection. That view can be stunning, however, as can be seen in the diamond ring and corona shots I managed to take with a 400mm telephoto.
Fair points about seeing what it expects, but then essentially all pattern recognition methods will need training data. CNNs need far more than most, which is why the authors need to resort to simulated data. However, suppose we have a method that needs fewer training data to reach the same performance, we would still have a method that realistically only finds objects similar to those examples fed to it. One way out of this problem is to devise methods that will flag strange outliers in the distribution of objects as "weird object, please let experts have a closer look". The data volumes are so vast that we will need methods to separate the masses of ordinary objects (for a given value of ordinary), and peculiar but well understood types, from the really weird objects that do not obviously fall into a known category. There are various efforts under way, including the EU SUNDIAL project. Exciting times.
I would have preferred they had settled for $5,000 plus legal costs, but I their method sets a very nice example
Precisely. I remember a lecture on "Ambient Intelligence" (anyone remember that catchphrase? Just one of the many phrases of things now going under the IoT moniker) at a conference, and the speaker raved on about how ideal it would be to have your home automatically start playing your favourite music when you entered. I suggested to the speaker that I (as many others) have rather wide-ranging tastes in music, and what I feel like depends HEAVILY on my mood. How would these things know what I wanted. The speaker gave a rather evasive answer and suggested errors weren't a huge problem, whereupon I suggested that if the system got it wrong and started playing the wrong music when I was in a particularly foul temper, I might go to its major data banks with a large axe and give it a reprogramming it would never forget. I added I also did not need refrigerators ordering beer for me, let alone self-satisfied doors, auto-chefs, or nutrimatic machines,
You can probably tell I had been playing some old HHGTTG tapes in the car on my way to the conference (as important educational material for the PhD students who were travelling with me)
Doffs hat (roo-leather Barmah today, it's pissing down) to the late, great Douglas Adams
If all goes according to plan, I should be in Glendo, Wyoming, right in the heart of the eclipse path, feverishly taking photos with a small telescope and a large telephoto. I managed to catch the 1999 eclipse just outside Sarreguemines in the Alsace, a grey day with just four perfectly chosen minutes of clear sky. I really hope I can improve on those (analogue) shots with the new digital kit. Wishing all watchers clear skies!
The moment the words "guinea pig" appeared, there was only one way it could end
cheering enthusiastically at the end of such a video aren't secretly a teensy bit disappointed the thing didn't go KABOOM!
Still, as a child of the Apollo era watching these videos, a wide grin always threatens to split my face at the ears, as so many great memories come flooding back
Good for NASA to get this test right!
(I recognise the difficulty in many other areas)
I am often hampered by not having access to code written by other scientists. If I want to test my shiny new algorithm against state-of-the-art efforts from others, it is a right pain to implement their methods from scratch. Besides, even if I write a working implementation (or get a student to do it), how can I be sure our implementation of the algorithm is the same as the original. I remember a Spanish colleague being very surprised at the timings we reported of his algorithm, because my student's implementation was a lot faster than his own, on a very similar machine. I also frequently just want to use a particular method in some image preprocessing step, and having some state-of-the-art method to work with directly saves a lot of time. People are getting much better in sharing academic code, but seeing a more concerted effort to retain access to code is useful.
The Cassini-Huygens mission just keeps on giving. I remember being thrilled by the descent on the Huygens probe onto Titan, and then there are all the data the Cassini orbiter has given. I love the way the scientists and engineers work tirelessly to get the most out of these missions, even from the final crash. Excellent example of international collaboration too
And my name isn't even Dave.
It might help, but the radiation isn't just coming through the lens, I would guess. Adding enough shielding to block all gamma rays might be impossible due to weight constraints. The X-rays blocked by CRT tubes are rather less energetic than these gamma rays, I think (10s of keV vs >100 keV, AFAIK)
Edit: I see I should speed-up my typing ;-)
Doffs hat to Steve Hersey!
I drained my cup of tea before reading this. Otherwise, my keyboard would have been in for another barrage of hot Keemun Congou tea ...
Great Friday read!!
From Ethereum via Ethereal to Ephemeral?
Not at all funny if you have been hit, of course.
See icon ->
I suppose I will have to make another cuppa
(the keyboard survived, no worries)
Very interesting data set. I might draw up some MSc thesis projects or student projects for the Computer Vision course I teach, developing methods to efficiently search for anomalies in those data. Much better than letting them use some toy data set. Not sure anything of note can be found, but at least there is some chance we can contribute a little
Alternatively, lack of sleep could be to blame. I recall Marvin's lullaby:
Now I lay me down in bed
Darkness won't enshroud my head
I can see in infrared
How I hate the night
Doffs hat (grey Tilley) to the late, great Douglas Adams once more
I bet it was all the self-satisfied doors that really got to him!
I suppose you deduced this from the discovery of a flattened coyote?
At ease, private Frazer!
Moist von Lipwig,
"... warned that any Bitcoins received after Monday, July 31, 2017 at GMT-0700 may vanish into thin air ..."
Bit like those very, very shiny gold coins some wizards (or indeed wizzards) would give you in payment
You could of course also use leaves as currency, although I think the last time that was used the official rate was several deciduous forests to one ship's peanut
My thoughts exactly. Digital watermarks are your friend here. Besides, AI may also be used to find the original, untampered footage, using content-based video retrieval, if said original footage is publicly available
THYGP XYWLK GHRNI QOPIY COVFE FETRU MPXPI
Tell the librarian he's a monkey, and see how that ends
Very interesting. More of a starlet than a star, but it does further constrain the border between red and brown dwarfs. I wonder if that boundary is influenced by helium content or metallicity. It is bound to, I suppose, given that the CNO cycle will depend on carbon, nitrogen, and oxygen (all metals to astronomers) abundances
Great boffinry. Looking forward to the results
Hard Rock Hotel hotel caught between a Rock and a Hard place
Sorry, just couldn't resist
Precisely. GPUs are great for SIMD-like parallelism, but it can be very difficult to harness when the task at hand does not allow such parallelism. In those cases, fewer, more complex and more independent compute cores can do a better job. Horses for courses, as ever. Many compute loads show a mixture of these SIMD and MIMD type parallel tasks, and for those the key factor in current designs is the fact that GPU and CPU memory is generally separate, and the speed of the bus linking the two is too low.
So, yes, GPUs are here to stay, but my (multi-core) CPU has this little piece of cardboard saying "I aten't dead!", and I think that is true.
The only reason I didn't spit tea all over the keyboard is that I had just finished my cup (an important precaution whenever I read the latest BOFH episode)
Wa can't dese stars stay in their own galaxy? We don't want no forrin stars! Takin' our jobs an going too fast
If it does, we might call it Eddie (assuming it has the GPP feature)
Copious amounts of data are easy to get, rather harder to turn into information. In order to train most AI or ML systems you need copious data with a reliable ground truth. The latter is very, very hard to come by, and requires lots of very, very careful, and usually dull work in labelling data items as belonging to different classes. If your ground truth on which you train you method is suspect, you will end up with over-fitting problems, because the ML/AI method with faithfully try to reproduce erroneous human decisions. For deep learning methods like convolutional neural networks (CNNs) to yield their (often impressive) results, you need hundreds of thousands, or preferably millions of accurately labelled data items. CNNs have been around fr quite a long time, but only the advent of large, labelled databases of images and the like made the methodology take off. labelling hundreds of thousands of data items automatically would be ideal, but isn't always possible. Usually some poor sods has to do lots and lots of unglamorous work.
Apart from these problems (which are daunting enough), there is the problem of all the parameter choices (learning rates, numbers and type of layers in deep networks, etc) to get right
Besides, what you learn in theory during lectures is not always what happens in reality. Modern HPC architectures are complex beast, far more complex than the Cray J932 on which I did my first serious HPC work: that was a shared memory set-up with no processor caches, and 32 identical vector processing CPUs. Predicting what is would do on different workloads was hard enough, but it is much harder with the current mix of multi-core CPU, GPU, and all manner of buses and caches working at different speeds that make up a modern cluster.
And anyway, this kind of competition is just fun!
Mainly for closure of the relatives of the victims
Interesting thought, which raises the question whether the column is vacuum, or merely vacuous? I would suggest the latter, as a true vacuum is so full of bovine excrement, which would exert a noticeable drag force on any sheep
Wouldn't Katie Hopkins' "thoughts" be anti-thoughts, so we have to look at anti-protons? (and yes, those have the (positive) same mass as protons, but it's the thought that counts)
Alternatively, given the often self-contradictory nature of what passes for thoughts in her case, some particle that is its own antiparticle (a Majorana fermion) would be suitable. As the most probable candidate for such a particle is the neutrino, which has near zero mass, this might be ideal
that when we find out exactly what all these constants are, the universe will instantly be replaced by something even more bizarrely inexplicable (maybe this has already happened)
Doffs hat (grey Tilley once more) to the late, great Douglas Adams
well, it is a refrigerator, after all
Oh dear, puns getting that bad this early in the week? Doesn't bode well, I'd better get me coat
Biting the hand that feeds IT © 1998–2017