That's the question. For strategy, I'd definitely go for Europa Universalis III (which is a much deeper game than the total war series, you can play hundreds of countries etc.) The lack of map flexibility isn't such a big deal when all the game scenarios are text-editable. Other possibilities would be Civ4 and GalCiv2.. The older games of either of those types have such easy AIs, replay-ability is compromised in my opinion.
For RPG, I'd probably take Angband, over Nethack by a small margin. Honor role would include Escape Velocity: Nova, NWN and Oblivion.
Of course, this is assuming no internet connection. With one, no question it'd be an MMORPG. I haven't played the most recent generation, so I'm not sure which I'd pick.
It's still very much news for nerds, stuff the matters. The the RIAA cases deal with an activity that many of us have participated in the past, present or future, and regardless of your opinion of the state of copyright law. It's important to know how one might respond if served.
More importantly though, these cases indirectly impact many other activities, many of which I think the great majority of the community feels are not unethical, which involve limits of copyrights, security, DRM, etc.
And speaking as somebody who's a US citizen who knows a fair amount as an amateur about the law and constitution as written documents, these cases, as well as the SCO entries, are certainly educating me greatly about how the law is actually practiced outside of my personal reading of it. Routine trial motions are relevant when they deal with something that is important to track, especially when most of the community doesn't know what's a routine motion and what isn't. I personally hope that at the very least everybody from the US learns from these, since being able to describe with accuracy and detail the problems we have with the current state of the laws is the only way that all the letters, e-mails etc to legislators are going to have any measurable impact.
If your house burns down, yes, you can go buy a new water heater, in which case you spend your own money. Or, if you had home insurance they'll replace the water heater, in which case all your local neighbors, friends and anybody vaguely connected with whatever stat bin they place you will also have to pay. Real goods have been stolen. They won't be replaced for free. Unlike many of the digital scams that are addressed on/., there is a real cost to every owner in the neighborhood, local or statistical, whether insurance pays the claim or not.
Hot water heater? Front door (if metal)? These are not cheap items.
I do think that the police should be the ones to handle this situation, and they'll certainly be able to get a warrant on the matter. However, if craigslist users think that this type of post is 'reasonable' to the point of pulling something as heavy as the heater, they better do something about dissuading people from making that type of post, whether it's public shaming, encouraging reporting of too-good-to-be-true posts or whatever else that somebody can dream up. In my opinion it's unreasonable for a classified service not to have some responsibility to guard against abuses such as what occured.
It's highly unsurprising that attention can be modulated by things we don't perceive. For a psychophysics study with an unbeatable title that supports the conclusion: " A gender- and sexual orientation-dependent spatial attentional effect of invisible images" http://www.pnas.org/cgi/content/abstract/060567810 3v1
I attended several elementary schools. The main one didn't give out homework until 6th grade (Bank Street, NYC), the school I went to in VT (Marion Cross) started giving homework in 3rd grade, the school in Berkeley (Cragmont) had homework in 1st grade, and I briefly went to a school in Bristol, England (Christ Church) that had minimal homework in 1st grade.
Of these schools, only Cragmont had heavy homework loads or emphasis at any point. I think that the problem with that, however is that I never formed the habit of doing homework, and still have difficulty just 'sitting down and doing work.' Homework outside of mathematics and reading is, IMHO of doubtful value until HS, and even then has limited utility. However, forming the habit of being able to sit down and do a set of work that needs to be done on your own time at home is highly useful throughout life.
Would they though? Izhikevich has taken a lot of time to try to get neurons into a 'reasonable' computational size, using a bunch of tricks from dynamical systems. This system may approach those dynamics, but it wasn't clear from the article. But that still isn't a general neuronal model. 'Regular' pyramidal cells often receive input from ~10-20k other cells, and there's no general description of which have an 'active' dendritic tree (ie. one that has areas that can spike towards the soma). There are plenty of other neurons, such as pyrimidal neurons in the Hippocampus, or Purkinje cells in the cerebellum, that we KNOW have active dendritic trees, and perform some pretty complex processing. And with a passive system, there's no reason for special processors, GPUs can do the computations just as well as any specialized chip (I know it isn't published yet, but check out things like http://cns.bu.edu/~elddm/ for examples of neural networks on GPUs).
I have to wonder what the purpose is.. You can model simplified 'point' neurons, and various aggregates that can be drawn from them (eg, McLoughlin's PDEs)... or you can run a simplified temporal dynamic (eg. Grossberg's 3D LAMINART), and easily include 200k+ neurons in the model easily to capture a broad range of function. For those would like running more detailed models of individual neuronal dynamics, you have Markram's project simulating a cortical column with compartmental models, or what Izhikevich is doing with delayed dynamic models.
Although this setup may be able to run ~1mil neurons, in total, it would seem that with 16 chips of 256x256 each, the level of interaction would be limited, and the article has no indication that these are the more complicated (and realistic) compartmental models of neurons that can sustain realistic individual neuronal dynamics (and for example Izhikevich, Markram and McLoughlin have spent a lot of time trying to simplify), or whether this is just running point style neurons a bit faster than is traditional.. and I have to wonder here, whether if these chips can't do compartmental models, why not just run this on a GPU?
I checked out this guy's webpage, and he seems smart.. but this project is years away from contributing.. I wonder, especially with the Poggio paper yesterday, when the best work being done just at MIT in Neuro/AI right now is probably in the Torralba lab, whether slashdot editors may want to find some people to vet the science submissions just a tad.
This is just from the first link on google searching "dollar value index" that's relevent (http://www.measuringworth.com/calculators/compare /result.php), so don't take this too strongly but..
(CD) In 1983, $33.36 from 1996 is worth:
$21.18 using the Consumer Price Index
$23.18 using the GDP deflator
$22.29 using the unskilled wage
$17.39 using the nominal GDP per capita
$15.09 using the relative share of GDP (Model T) In 1996, $890.00 from 1908 is worth:
$15,654.82 using the Consumer Price Index
$12,087.77 using the GDP deflator
$64,267.11 using the unskilled wage
$75,533.01 using the nominal GDP per capita
$229,378.21 using the relative share of GDP
Looks to me like their both overestimating both the relative price of the CD, and also not considereing the general increase in income compared to the rest of the world that we've gone through in the interim, see the model T price.
I wonder though, whether EULAs couldn't also be used as a defense against this type of trolling in a way patents can't. Suppose, for example, that a maker of a small, free-ware (popular) program, put a clause into their EULA saying in essence "You agree not sue any other end-users of this product".. if there were enough of these out there, any EULA-troll company would have almost certainly have clicked through on one of them, which would put them in a position of arguing both for and against EULA terms.
Also, if the above wouldn't have legal standing, I wonder if software companies might put no-action clauses in all of their software, so any litigation over EULAs they agreed to might also have to consider the EULAs the trolls clicked through on their accounting software, etc.
These issues obviously get very sticky and complicated extremely quickly, which highlights the absurdity of the entire EULA as contract idea.
The basic issue of porn, etc, isn't gonna go away: a significant proportion of people think that sex is bad/dirty etc. In the US we now have a fairly zealous set of laws prohibiting various sexual action/production (just look at the ESPN.com headline yesterday about the 17 year old who's in prison 10 year mandatory for getting a blow job from a 15 year old). With people that are willing to agitate for these beliefs around, I think in terms of technology we should work to make things like porn as clearly classed as possible, like the xxx domain. I would much rather fight over these issues in the.xxx domain, rather than having my freedoms circumscribed in misguided efforts to attain the approval of the zealous because porn is 'hard to filter.'
In all seriousness, going to the manufacturer's website is a great idea if they have a support forum. No quicker way to get a feel for the bugs and problems you may encounter.
I call BS! As a scientist who works in spatial attention, I'm not aware of any consensus or studies that are well accepted showing general male superiority in spatial processing (I'm not sufficiently familiar with the language literature to address whether females might generally have better language facilities). It is true that several studies have shown males tend to have larger interparietal lobules (IPLs) than females (Frederikse et al., 1999), and that there is some evidence suggesting a degree of right lateralization in processing of spatial attention. But more recent fMRI studies that have operated at higher statistical thresholds have cast doubt on right lateralization of attentional function, and the IPL is one of a whole group of areas (interparietal sulcus, cingulate, dorsal lateral pre-frontal cortex, etc) that are implicated in processing complex spatial tasks (too many studies to list, but generally looking up people like Sereno or Kanwisher will get you started).
Beyond that, with evidence that training attention, for example playing FPS games (probably a popular method here on/.), can have massive effects on performance in attention demanding tasks (Green & Bavelier, 2003). Training evidence combined with the huge individual variation in individual performance, makes it utter trash to say women don't have the same 'potential' in science or engineering as their male counterparts.
First of all, I'd like to point people to: http://www.fieldofschemes.com/ which details how sports teams use public money. Although the editorial is certainly against stadiums, the numbers are about the best you can find.
Since I've been following the A's stadium on the site mentioned above for over a year, I can tell you that it is by no means a done deal. Among other things, there aren't enough police to regulate games, and who's to pay for the increase necessary for that is absent in the current deal.
Classification systems such as this are still fundementally in an early form of development. The purpose of these things, whether computer vision or patient classifiers to aid in medical diagnosis is not to replace the doctor/human observer. It is either to pre-screen, or provide a second opinion that may cause further testing to be done. This system would work just fine if, when it detected violent behavior, it flagged it to the attention of a human observer who could then evaluate whether it was a false positive or not. But to go immediately from 'Computer Vision detected possible violent behavior' to 'Send a police car' would be farcical.
If these cameras were controlled in such a way that a human observer only ever saw things that were classified as violent, rather than the human observer being able to look at all the camera outputs at any given time, it would protect our privacy more than current surveillence systems, since there would be no surveillence without some modicum of cause.
No, but you could probably patent methods for lobbying congress for tax increases or decreases in the interest of specific clients, and if you got it you might be able to shutdown a lot of lobbying efforts.
That is totally untrue. The entire MEDLINE database, nearly all of Science Web (isiknowledge.com) is PDF. There have been millions of hours spent creating and indexing much our science today in PDF files. There ain't gonna be a quick changeover. Most scientists are inherently conservative about things like this, because not unreasonably, they assume that any new standard is going to screw their previous databases. A large proportion of the publications in this country relies on federal grant money, and both the grants and all that has been published as results are in PDF.
PDF, after over a decade in existence has gained a standard foothold in a wide variety of fields, anybody who believes that there's gonna be a second change in less than that time needs to make a reality check.
I think the main reason that so few kids today get into coding is that computers have become so useable for most tasks, that it's no longer necessary to program much yourself, too many things already exsist. For those of us who grew up in the 80s or 90s, home computers were new, tended to be kludgy and have a minimum of litter gadget type programs to do things that you wanted. The fastest way to create something like a calculator in dos was to write it in basic (and it was a fun exercise). Likewise, I spent many hours doing logo-writer programming on various projects (often for solving various geometry things) that I had. Other things that today we take for granted, like for example saved game editors, were unavailable unless you were able to get a hex editor and systematically edit the game looking for changes.
These days, nearly all those things are included in the OS or easily available via web searching. There's no motivation to code yourself when the solution is a 10 minute google search away. The easy availablity and convenience of the web, and in fact the open source movement with all of the 80s and 90s tinkerer generation publishing their tinkering has removed that motivation that we all had to code in the first place for the next generation. While I agree that there isn't something as compelling and easy as logowriter out there today to get started on, I think the underlying cause is much more based on the sociology.
Your sleep cycles are regulated by two main features: environment (most importantly exposure to bright light when you get up), and circadian rhythms, which are a natural ~25hr rhythm generated in the suprachiasmatic nucleus of your hypothalmus (SCN), which is receives light signals and uses them to entrain your rhythms to a 24 hour day. The centers for controlling your sleep function mostly reside in the brainstem, although other areas such as parts of the thalmus and the pineal gland play a role.. all these areas appear to combine to function more like an on-off switch. The interaction of this switch and your circadian rhythms control whether you're asleep and how asleep you are. I should also note, that this switch is messy, and in a typical nights sleep, you'll have a fat tailed distribution of sleep length (ie. 100's of times asleep for a few seconds, once asleep for a few hours). The switching in and out is thought to modulate what sleep 'phase' you're in, which controls such things as the calcium channels in your thalmus (which can drive neurons when you're asleep but not when you're awake), whether you're undergoing REM and the strength of theta rhythms in your hippocampus, which appears vital in your ability to consolidate long term memory.
Wake up grogginess is much more correlated with the sleep phase you wake up out of, whether you were in REM, whether you're being roused from a long or short bout of actual sleep and how close the wake-up time is to your current circadian entrainment. Your getting online faster in the morning is much more related to how strongly entrained your circadian rhythms are than having a fast or slow thalmus--note that your description of waking up with the sun is the exact time that's most advantageous to use the sun to entrain your circadian rhythms. I should note on cortical processing 'speed' most neurons are contrained by their RC constants (neurons are capacitors), which in most of cortex is typically limited to about 10ms (areas of the auditory system in particular, get these down to near 1ms using small cell bodies, specially maintained voltage differences that are energy expensive, etc). This is not something subject to interpersonal variance, it's basically the same in both humans and frogs.
This paper's finding, is that ethanol can block T-type calcium channels turning on, which is thought to be vital in things like having a strong theta rhythm while asleep (which would, for example, be a reason why you have a really hard time remembering what you did the night before if you were fairly drunk, because your theta rhythms would have been compromised, inhibiting your ability to consolidate long term memory).
I was a math/history double major, and am now doing neuroscience... but that's besides the point.
With a pure math BA you can basically go to any engineering, physics, biology, neuroscience, finance, econ, cs, etc masters or PhD program and do just fine. The important part about a math degree, is that it gives you the background and experience required to learn specific applications really quickly. There's a huge demand out there for people who are talented at math, although most of this demand isn't 'pure' math per se, there are a lot of interesting applied problems you can work on that do have theoretical interest to a mathematician.
You should really have no problem finding a job or getting into grad school in almost any tech/science type field that you're interested in coming out college with a BA in Math. The great thing about a math major, against a more specific applied major, is that you learn how to think about many of the applied problems in a deeper way, and since you're aquainted with the underlying theory, you can much more easily link various ideas that are only taught at a plug and play level in the applied fields (for example, most IOE curriculum is just rather narrow subset of graph theory & combinatorics).
Personally, I was interested in a lot of things as an undergrad, and decided to major in math since it basically kept all my options open on a grad/job level, and I certainly haven't regretted that decision.
1) Work in a cool & dry environment. If it's too humid or warm, you'll get lethargic and have trouble getting motivated.
2) When you're sitting in front of the computer and the work is just not happening, and you've reloaded slashdot 5 times in the last 5 minutes etc, get up and take a break. Don't force yourself to sit there. Walk around the room, or better yet that's a good time to walk to the store and do errands.
3) On the similar vein to #2, get yourself on a normal exercise program. Not a beat-the-crap-outta-yourself one, but something like trying to bike for an hour a day. When you're not motivated to work, get on the bike and go work out. It'll both give your mind a chance to float (and often come up with an idea about what you're working on), but once you make this a habit, you'll be more invigorated when you're done.
4) Most people can work most efficiently on stuff like programming immediately after they wake up (whenever that is;p). I find that if immediately after I wake up, I sit down in front of the computer and work until I'm bored it really makes me get a lot more work done. On the other hand, if I sit down in front of the computer and start checking news sites, etc. then it often takes me a lot longer to initiate doing the real work.. which is the major struggle in the first place.
5) In terms of diet, caffeine and stuff like that.. I personally find that I work best when I'm not drinking, not eating sugary things that'll cause my blood sugar to fluctuate, and not on caffeine unless I absolutely need to stay up because a deadline is looming (another advantage of that is the caffeine works a lot better).
But beyond that, I'd echo other people's comments that if you're not having a problem supporting your lifestyle workings 25hrs/week, then take advantage of that and do things you find fun, indulge your hobbies etc. Chances are you'll wind up being forced into 40+hr/week situations for periods in the future, so take advantage of your current situation.
One of the basic tasks our visual system is much, much better at executing than computers is visual search. The basic 'experiment' is that you are asked either a question like "Is there a red car in this picture?" (natural images) or "Are all the lines the same orientation?" (more traditional psychophysics). Then images are displayed, and our response time is recorded. Early experiments in the visual search paradigm appeared to show that there was two classes of search stimuli: those that 'pop-out' and those that require incremental search. The difference is that in pop-out conditions, increasing the number of elements in the image does not increase search time, while in incremental it does at XXms/element... and generally it takes about twice as long for us to respond if there is no positive element.
One main theory on how our brain does this, Feature Integration Theory by Anne Treisman (or similar but more recent, Guided Search by Jeremy Wolfe), which many computer vision algorithms try to copy, asserts that there are various feature maps for certain quantities like color, orientation, depth, spatial scale, etc. These are combined into a saliency map which is a weighted average of the feature maps. Things pop-out when the target has high salience compare to the background, for example it's easy to find the red T in a background of blue T's, but not so easy to find the red L in a background of red T's and blue L's.
Now, it appears from the article, and what little they say on the Lab webpage, that they are trying to measure EEG responses (which are quite crude) during rapid serial search tasks, in order to prime a computer vision object recognition system, which is then only run on those images human's appear to find sufficintly salient when they see them. This saves the time of a person actually having to search and make a decision about an image, while utilizing the visual systems incredibly powerful early 'pre-attentive' form & object binding resources.
If there is a sufficiently high signal from the EEG to do that after say, 100ms display times, then I think this could be useful for certain types of search task. However, due to the time courses present in most visual search experiments, the fact that it's not totally apparent how efficient certain parts of our saliency system actually are (check our Jeremy Wolfe's reviews for more data), I'm totally unconvinced that this type of system will give you a sufficent signal to noise ratio to be worth using for anything. This is especially true because of another perceptual phenomenon in search, which is that your error rate basically shoots up exponentially as the probability of a positive goes down. This is to say, in an experiment where a normal observer would have a 99% accuracy rate with 50% of the images containing the target, this drops to 60% accuracy for 10% target positive, and only 30% accuracy at 1% target positive (numbers fudged, but ballpark, since I'm too lazy to look them up). If this has its roots in insufficient priming in early vision, for example, then this entire scheme flops just as badly as using a human for tasks like finding the bomb in the x-ray image of the suitcase... and we haven't even started to get into issues of the person not actually looking at the image because they're bored, etc.
As it is, DARPA is spending a mere 758k, which is chump change for them, and there's a decent chance that it'll work in certain specific but useful circumstances which may warrant the research.
I won't rehash posts about how far fMRI is away from being a useful measuring device in this regard, since deviations are generally small enough that measurements over many trials must be aggregated to achieve significant results in carefully controlled conditions. But even if fMRI's were much better, and we understood how the brain worked much more closely this would still be of limited to no usefulness as an actual scientific method (it would probably would better than the polygraph, but would still be pseudoscience).
The problem is that the polygraph works in this basic manner: The examiner asks you a whole bunch of filler questions, claiming these are 'controls.' These results are all ignored. Questions in this phase are things like "Is today Tuesday?". Then the examiner intersperses the real controls (he's already lied to you about what they are), questions which they'll preface with ominous portents if you answer affirmatively, so the examiner assumes you're going to lie about them ("Have you ever cheated on a girlfriend? Have you ever used marijuana?).
Then the examiner takes the second controls and compares them to his test questions. If you're test questions exceed the response from the (presumed to be lying) controls, the examiner assumes you're lying. Thus, telling the truth throughout the entire procedure is liable to land you in hot water. (For more information, from an admittedley 'biased' site, but I think they're pretty clear can be found at http://www.antipolygraph.org/).
However, a true lie detector test would require a much more coherent defintion of what a lie is, which is very hard to create. Most people would agree that actively misleading somebody with no regard to your factual knowledge is lying. This also tends to be a useless type of lie in these situations because people get there stories mixed up, or they don't think through all the details. Much more common types of lies, are witholding useful information while truthfully relating aspects of the response, or changing the context of the answer, and other things which mislead but do not show complete disregard for the truth. The best lies in the intelligence useful/lessness sense are those that only minorly distort the truth, but in a particularly significant way.
Until you can metrize all these different types of not being truthful, or of avoiding certain facts etc, and until you can metrize their reponses for each individual (my guess is that this type of thing will have a high variance between people), you can't produce anything that can reasonably be called a scientific lie detector.
Slashdot Mirror
That's the question. For strategy, I'd definitely go for Europa Universalis III (which is a much deeper game than the total war series, you can play hundreds of countries etc.) The lack of map flexibility isn't such a big deal when all the game scenarios are text-editable. Other possibilities would be Civ4 and GalCiv2.. The older games of either of those types have such easy AIs, replay-ability is compromised in my opinion.
For RPG, I'd probably take Angband, over Nethack by a small margin. Honor role would include Escape Velocity: Nova, NWN and Oblivion.
Of course, this is assuming no internet connection. With one, no question it'd be an MMORPG. I haven't played the most recent generation, so I'm not sure which I'd pick.
It's still very much news for nerds, stuff the matters. The the RIAA cases deal with an activity that many of us have participated in the past, present or future, and regardless of your opinion of the state of copyright law. It's important to know how one might respond if served.
More importantly though, these cases indirectly impact many other activities, many of which I think the great majority of the community feels are not unethical, which involve limits of copyrights, security, DRM, etc.
And speaking as somebody who's a US citizen who knows a fair amount as an amateur about the law and constitution as written documents, these cases, as well as the SCO entries, are certainly educating me greatly about how the law is actually practiced outside of my personal reading of it. Routine trial motions are relevant when they deal with something that is important to track, especially when most of the community doesn't know what's a routine motion and what isn't. I personally hope that at the very least everybody from the US learns from these, since being able to describe with accuracy and detail the problems we have with the current state of the laws is the only way that all the letters, e-mails etc to legislators are going to have any measurable impact.
If your house burns down, yes, you can go buy a new water heater, in which case you spend your own money. Or, if you had home insurance they'll replace the water heater, in which case all your local neighbors, friends and anybody vaguely connected with whatever stat bin they place you will also have to pay. Real goods have been stolen. They won't be replaced for free. Unlike many of the digital scams that are addressed on /., there is a real cost to every owner in the neighborhood, local or statistical, whether insurance pays the claim or not.
Hot water heater? Front door (if metal)? These are not cheap items.
I do think that the police should be the ones to handle this situation, and they'll certainly be able to get a warrant on the matter. However, if craigslist users think that this type of post is 'reasonable' to the point of pulling something as heavy as the heater, they better do something about dissuading people from making that type of post, whether it's public shaming, encouraging reporting of too-good-to-be-true posts or whatever else that somebody can dream up. In my opinion it's unreasonable for a classified service not to have some responsibility to guard against abuses such as what occured.
See:/ 16/208250
http://politics.slashdot.org/article.pl?sid=06/11
or
http://www.slate.com/id/2153352/?nav=tap3
It's highly unsurprising that attention can be modulated by things we don't perceive. For a psychophysics study with an unbeatable title that supports the conclusion: " A gender- and sexual orientation-dependent spatial attentional effect of invisible images" http://www.pnas.org/cgi/content/abstract/060567810 3v1
I attended several elementary schools. The main one didn't give out homework until 6th grade (Bank Street, NYC), the school I went to in VT (Marion Cross) started giving homework in 3rd grade, the school in Berkeley (Cragmont) had homework in 1st grade, and I briefly went to a school in Bristol, England (Christ Church) that had minimal homework in 1st grade.
Of these schools, only Cragmont had heavy homework loads or emphasis at any point. I think that the problem with that, however is that I never formed the habit of doing homework, and still have difficulty just 'sitting down and doing work.' Homework outside of mathematics and reading is, IMHO of doubtful value until HS, and even then has limited utility. However, forming the habit of being able to sit down and do a set of work that needs to be done on your own time at home is highly useful throughout life.
Would they though? Izhikevich has taken a lot of time to try to get neurons into a 'reasonable' computational size, using a bunch of tricks from dynamical systems. This system may approach those dynamics, but it wasn't clear from the article. But that still isn't a general neuronal model. 'Regular' pyramidal cells often receive input from ~10-20k other cells, and there's no general description of which have an 'active' dendritic tree (ie. one that has areas that can spike towards the soma). There are plenty of other neurons, such as pyrimidal neurons in the Hippocampus, or Purkinje cells in the cerebellum, that we KNOW have active dendritic trees, and perform some pretty complex processing. And with a passive system, there's no reason for special processors, GPUs can do the computations just as well as any specialized chip (I know it isn't published yet, but check out things like http://cns.bu.edu/~elddm/ for examples of neural networks on GPUs).
I have to wonder what the purpose is.. You can model simplified 'point' neurons, and various aggregates that can be drawn from them (eg, McLoughlin's PDEs)... or you can run a simplified temporal dynamic (eg. Grossberg's 3D LAMINART), and easily include 200k+ neurons in the model easily to capture a broad range of function. For those would like running more detailed models of individual neuronal dynamics, you have Markram's project simulating a cortical column with compartmental models, or what Izhikevich is doing with delayed dynamic models.
Although this setup may be able to run ~1mil neurons, in total, it would seem that with 16 chips of 256x256 each, the level of interaction would be limited, and the article has no indication that these are the more complicated (and realistic) compartmental models of neurons that can sustain realistic individual neuronal dynamics (and for example Izhikevich, Markram and McLoughlin have spent a lot of time trying to simplify), or whether this is just running point style neurons a bit faster than is traditional.. and I have to wonder here, whether if these chips can't do compartmental models, why not just run this on a GPU?
I checked out this guy's webpage, and he seems smart.. but this project is years away from contributing.. I wonder, especially with the Poggio paper yesterday, when the best work being done just at MIT in Neuro/AI right now is probably in the Torralba lab, whether slashdot editors may want to find some people to vet the science submissions just a tad.
This is just from the first link on google searching "dollar value index" that's relevent (http://www.measuringworth.com/calculators/compare /result.php), so don't take this too strongly but..
(CD)
In 1983, $33.36 from 1996 is worth:
$21.18 using the Consumer Price Index
$23.18 using the GDP deflator
$22.29 using the unskilled wage
$17.39 using the nominal GDP per capita
$15.09 using the relative share of GDP
(Model T)
In 1996, $890.00 from 1908 is worth:
$15,654.82 using the Consumer Price Index
$12,087.77 using the GDP deflator
$64,267.11 using the unskilled wage
$75,533.01 using the nominal GDP per capita
$229,378.21 using the relative share of GDP
Looks to me like their both overestimating both the relative price of the CD, and also not considereing the general increase in income compared to the rest of the world that we've gone through in the interim, see the model T price.
I wonder though, whether EULAs couldn't also be used as a defense against this type of trolling in a way patents can't. Suppose, for example, that a maker of a small, free-ware (popular) program, put a clause into their EULA saying in essence "You agree not sue any other end-users of this product".. if there were enough of these out there, any EULA-troll company would have almost certainly have clicked through on one of them, which would put them in a position of arguing both for and against EULA terms.
Also, if the above wouldn't have legal standing, I wonder if software companies might put no-action clauses in all of their software, so any litigation over EULAs they agreed to might also have to consider the EULAs the trolls clicked through on their accounting software, etc.
These issues obviously get very sticky and complicated extremely quickly, which highlights the absurdity of the entire EULA as contract idea.
The basic issue of porn, etc, isn't gonna go away: a significant proportion of people think that sex is bad/dirty etc. In the US we now have a fairly zealous set of laws prohibiting various sexual action/production (just look at the ESPN.com headline yesterday about the 17 year old who's in prison 10 year mandatory for getting a blow job from a 15 year old). With people that are willing to agitate for these beliefs around, I think in terms of technology we should work to make things like porn as clearly classed as possible, like the xxx domain. I would much rather fight over these issues in the .xxx domain, rather than having my freedoms circumscribed in misguided efforts to attain the approval of the zealous because porn is 'hard to filter.'
In all seriousness, going to the manufacturer's website is a great idea if they have a support forum. No quicker way to get a feel for the bugs and problems you may encounter.
I call BS! As a scientist who works in spatial attention, I'm not aware of any consensus or studies that are well accepted showing general male superiority in spatial processing (I'm not sufficiently familiar with the language literature to address whether females might generally have better language facilities). It is true that several studies have shown males tend to have larger interparietal lobules (IPLs) than females (Frederikse et al., 1999), and that there is some evidence suggesting a degree of right lateralization in processing of spatial attention. But more recent fMRI studies that have operated at higher statistical thresholds have cast doubt on right lateralization of attentional function, and the IPL is one of a whole group of areas (interparietal sulcus, cingulate, dorsal lateral pre-frontal cortex, etc) that are implicated in processing complex spatial tasks (too many studies to list, but generally looking up people like Sereno or Kanwisher will get you started).
/.), can have massive effects on performance in attention demanding tasks (Green & Bavelier, 2003). Training evidence combined with the huge individual variation in individual performance, makes it utter trash to say women don't have the same 'potential' in science or engineering as their male counterparts.
Beyond that, with evidence that training attention, for example playing FPS games (probably a popular method here on
First of all, I'd like to point people to: http://www.fieldofschemes.com/ which details how sports teams use public money. Although the editorial is certainly against stadiums, the numbers are about the best you can find.
Since I've been following the A's stadium on the site mentioned above for over a year, I can tell you that it is by no means a done deal. Among other things, there aren't enough police to regulate games, and who's to pay for the increase necessary for that is absent in the current deal.
Classification systems such as this are still fundementally in an early form of development. The purpose of these things, whether computer vision or patient classifiers to aid in medical diagnosis is not to replace the doctor/human observer. It is either to pre-screen, or provide a second opinion that may cause further testing to be done. This system would work just fine if, when it detected violent behavior, it flagged it to the attention of a human observer who could then evaluate whether it was a false positive or not. But to go immediately from 'Computer Vision detected possible violent behavior' to 'Send a police car' would be farcical.
If these cameras were controlled in such a way that a human observer only ever saw things that were classified as violent, rather than the human observer being able to look at all the camera outputs at any given time, it would protect our privacy more than current surveillence systems, since there would be no surveillence without some modicum of cause.
No, but you could probably patent methods for lobbying congress for tax increases or decreases in the interest of specific clients, and if you got it you might be able to shutdown a lot of lobbying efforts.
That is totally untrue. The entire MEDLINE database, nearly all of Science Web (isiknowledge.com) is PDF. There have been millions of hours spent creating and indexing much our science today in PDF files. There ain't gonna be a quick changeover. Most scientists are inherently conservative about things like this, because not unreasonably, they assume that any new standard is going to screw their previous databases. A large proportion of the publications in this country relies on federal grant money, and both the grants and all that has been published as results are in PDF.
PDF, after over a decade in existence has gained a standard foothold in a wide variety of fields, anybody who believes that there's gonna be a second change in less than that time needs to make a reality check.
I think the main reason that so few kids today get into coding is that computers have become so useable for most tasks, that it's no longer necessary to program much yourself, too many things already exsist. For those of us who grew up in the 80s or 90s, home computers were new, tended to be kludgy and have a minimum of litter gadget type programs to do things that you wanted. The fastest way to create something like a calculator in dos was to write it in basic (and it was a fun exercise). Likewise, I spent many hours doing logo-writer programming on various projects (often for solving various geometry things) that I had. Other things that today we take for granted, like for example saved game editors, were unavailable unless you were able to get a hex editor and systematically edit the game looking for changes.
These days, nearly all those things are included in the OS or easily available via web searching. There's no motivation to code yourself when the solution is a 10 minute google search away. The easy availablity and convenience of the web, and in fact the open source movement with all of the 80s and 90s tinkerer generation publishing their tinkering has removed that motivation that we all had to code in the first place for the next generation. While I agree that there isn't something as compelling and easy as logowriter out there today to get started on, I think the underlying cause is much more based on the sociology.
Short answer is no.
Your sleep cycles are regulated by two main features: environment (most importantly exposure to bright light when you get up), and circadian rhythms, which are a natural ~25hr rhythm generated in the suprachiasmatic nucleus of your hypothalmus (SCN), which is receives light signals and uses them to entrain your rhythms to a 24 hour day. The centers for controlling your sleep function mostly reside in the brainstem, although other areas such as parts of the thalmus and the pineal gland play a role.. all these areas appear to combine to function more like an on-off switch. The interaction of this switch and your circadian rhythms control whether you're asleep and how asleep you are. I should also note, that this switch is messy, and in a typical nights sleep, you'll have a fat tailed distribution of sleep length (ie. 100's of times asleep for a few seconds, once asleep for a few hours). The switching in and out is thought to modulate what sleep 'phase' you're in, which controls such things as the calcium channels in your thalmus (which can drive neurons when you're asleep but not when you're awake), whether you're undergoing REM and the strength of theta rhythms in your hippocampus, which appears vital in your ability to consolidate long term memory.
Wake up grogginess is much more correlated with the sleep phase you wake up out of, whether you were in REM, whether you're being roused from a long or short bout of actual sleep and how close the wake-up time is to your current circadian entrainment. Your getting online faster in the morning is much more related to how strongly entrained your circadian rhythms are than having a fast or slow thalmus--note that your description of waking up with the sun is the exact time that's most advantageous to use the sun to entrain your circadian rhythms. I should note on cortical processing 'speed' most neurons are contrained by their RC constants
(neurons are capacitors), which in most of cortex is typically limited to about 10ms (areas of the auditory system in particular, get these down to near 1ms using small cell bodies, specially maintained voltage differences that are energy expensive, etc). This is not something subject to interpersonal variance, it's basically the same in both humans and frogs.
This paper's finding, is that ethanol can block T-type calcium channels turning on, which is thought to be vital in things like having a strong theta rhythm while asleep (which would, for example, be a reason why you have a really hard time remembering what you did the night before if you were fairly drunk, because your theta rhythms would have been compromised, inhibiting your ability to consolidate long term memory).
As long as you're willing to rat out the reseller, MS will replace your pirated version with a genuine one.
U sers_Get_Amnesty/1115239342
http://www.betanews.com/article/NonLegit_Windows_
I was a math/history double major, and am now doing neuroscience... but that's besides the point.
With a pure math BA you can basically go to any engineering, physics, biology, neuroscience, finance, econ, cs, etc masters or PhD program and do just fine. The important part about a math degree, is that it gives you the background and experience required to learn specific applications really quickly. There's a huge demand out there for people who are talented at math, although most of this demand isn't 'pure' math per se, there are a lot of interesting applied problems you can work on that do have theoretical interest to a mathematician.
You should really have no problem finding a job or getting into grad school in almost any tech/science type field that you're interested in coming out college with a BA in Math. The great thing about a math major, against a more specific applied major, is that you learn how to think about many of the applied problems in a deeper way, and since you're aquainted with the underlying theory, you can much more easily link various ideas that are only taught at a plug and play level in the applied fields (for example, most IOE curriculum is just rather narrow subset of graph theory & combinatorics).
Personally, I was interested in a lot of things as an undergrad, and decided to major in math since it basically kept all my options open on a grad/job level, and I certainly haven't regretted that decision.
1) Work in a cool & dry environment. If it's too humid or warm, you'll get lethargic and have trouble getting motivated.
;p). I find that if immediately after I wake up, I sit down in front of the computer and work until I'm bored it really makes me get a lot more work done. On the other hand, if I sit down in front of the computer and start checking news sites, etc. then it often takes me a lot longer to initiate doing the real work.. which is the major struggle in the first place.
2) When you're sitting in front of the computer and the work is just not happening, and you've reloaded slashdot 5 times in the last 5 minutes etc, get up and take a break. Don't force yourself to sit there. Walk around the room, or better yet that's a good time to walk to the store and do errands.
3) On the similar vein to #2, get yourself on a normal exercise program. Not a beat-the-crap-outta-yourself one, but something like trying to bike for an hour a day. When you're not motivated to work, get on the bike and go work out. It'll both give your mind a chance to float (and often come up with an idea about what you're working on), but once you make this a habit, you'll be more invigorated when you're done.
4) Most people can work most efficiently on stuff like programming immediately after they wake up (whenever that is
5) In terms of diet, caffeine and stuff like that.. I personally find that I work best when I'm not drinking, not eating sugary things that'll cause my blood sugar to fluctuate, and not on caffeine unless I absolutely need to stay up because a deadline is looming (another advantage of that is the caffeine works a lot better).
But beyond that, I'd echo other people's comments that if you're not having a problem supporting your lifestyle workings 25hrs/week, then take advantage of that and do things you find fun, indulge your hobbies etc. Chances are you'll wind up being forced into 40+hr/week situations for periods in the future, so take advantage of your current situation.
One of the basic tasks our visual system is much, much better at executing than computers is visual search. The basic 'experiment' is that you are asked either a question like "Is there a red car in this picture?" (natural images) or "Are all the lines the same orientation?" (more traditional psychophysics). Then images are displayed, and our response time is recorded. Early experiments in the visual search paradigm appeared to show that there was two classes of search stimuli: those that 'pop-out' and those that require incremental search. The difference is that in pop-out conditions, increasing the number of elements in the image does not increase search time, while in incremental it does at XXms/element... and generally it takes about twice as long for us to respond if there is no positive element.
One main theory on how our brain does this, Feature Integration Theory by Anne Treisman (or similar but more recent, Guided Search by Jeremy Wolfe), which many computer vision algorithms try to copy, asserts that there are various feature maps for certain quantities like color, orientation, depth, spatial scale, etc. These are combined into a saliency map which is a weighted average of the feature maps. Things pop-out when the target has high salience compare to the background, for example it's easy to find the red T in a background of blue T's, but not so easy to find the red L in a background of red T's and blue L's.
Now, it appears from the article, and what little they say on the Lab webpage, that they are trying to measure EEG responses (which are quite crude) during rapid serial search tasks, in order to prime a computer vision object recognition system, which is then only run on those images human's appear to find sufficintly salient when they see them. This saves the time of a person actually having to search and make a decision about an image, while utilizing the visual systems incredibly powerful early 'pre-attentive' form & object binding resources.
If there is a sufficiently high signal from the EEG to do that after say, 100ms display times, then I think this could be useful for certain types of search task. However, due to the time courses present in most visual search experiments, the fact that it's not totally apparent how efficient certain parts of our saliency system actually are (check our Jeremy Wolfe's reviews for more data), I'm totally unconvinced that this type of system will give you a sufficent signal to noise ratio to be worth using for anything. This is especially true because of another perceptual phenomenon in search, which is that your error rate basically shoots up exponentially as the probability of a positive goes down. This is to say, in an experiment where a normal observer would have a 99% accuracy rate with 50% of the images containing the target, this drops to 60% accuracy for 10% target positive, and only 30% accuracy at 1% target positive (numbers fudged, but ballpark, since I'm too lazy to look them up). If this has its roots in insufficient priming in early vision, for example, then this entire scheme flops just as badly as using a human for tasks like finding the bomb in the x-ray image of the suitcase... and we haven't even started to get into issues of the person not actually looking at the image because they're bored, etc.
As it is, DARPA is spending a mere 758k, which is chump change for them, and there's a decent chance that it'll work in certain specific but useful circumstances which may warrant the research.
I won't rehash posts about how far fMRI is away from being a useful measuring device in this regard, since deviations are generally small enough that measurements over many trials must be aggregated to achieve significant results in carefully controlled conditions. But even if fMRI's were much better, and we understood how the brain worked much more closely this would still be of limited to no usefulness as an actual scientific method (it would probably would better than the polygraph, but would still be pseudoscience).
The problem is that the polygraph works in this basic manner:
The examiner asks you a whole bunch of filler questions, claiming these are 'controls.' These results are all ignored. Questions in this phase are things like "Is today Tuesday?". Then the examiner intersperses the real controls (he's already lied to you about what they are), questions which they'll preface with ominous portents if you answer affirmatively, so the examiner assumes you're going to lie about them ("Have you ever cheated on a girlfriend? Have you ever used marijuana?).
Then the examiner takes the second controls and compares them to his test questions. If you're test questions exceed the response from the (presumed to be lying) controls, the examiner assumes you're lying. Thus, telling the truth throughout the entire procedure is liable to land you in hot water. (For more information, from an admittedley 'biased' site, but I think they're pretty clear can be found at http://www.antipolygraph.org/).
However, a true lie detector test would require a much more coherent defintion of what a lie is, which is very hard to create. Most people would agree that actively misleading somebody with no regard to your factual knowledge is lying. This also tends to be a useless type of lie in these situations because people get there stories mixed up, or they don't think through all the details. Much more common types of lies, are witholding useful information while truthfully relating aspects of the response, or changing the context of the answer, and other things which mislead but do not show complete disregard for the truth. The best lies in the intelligence useful/lessness sense are those that only minorly distort the truth, but in a particularly significant way.
Until you can metrize all these different types of not being truthful, or of avoiding certain facts etc, and until you can metrize their reponses for each individual (my guess is that this type of thing will have a high variance between people), you can't produce anything that can reasonably be called a scientific lie detector.