Huliwuxian

Notes: 5 / 1 hour ago from spaceships
David Brooks crapping stuff out of his ass, as per usual
spaceships:

Instead of reducing costs, the bills in Congress would probably raise them. They would mean that more of the nation’s wealth would be siphoned off from productive uses and shifted into a still wasteful health care system.

… The bottom line is that we face a brutal choice.

Reform would make us a more decent society, but also a less vibrant one. It would ease the anxiety of millions at the cost of future growth. It would heal a wound in the social fabric while piling another expensive and untouchable promise on top of the many such promises we’ve already made. America would be a less youthful, ragged and unforgiving nation, and a more middle-aged, civilized and sedate one.

We all have to decide what we want at this moment in history, vitality or security. We can debate this or that provision, but where we come down will depend on that moral preference. Don’t get stupefied by technical details. This debate is about values.

Notice that the last paragraph is just a bunch of random words strung together without any logical progression.

All these people crying about the ballooning deficit and bankrupting the country and uncontrolled spending and blah blah blah, where the fuck were you in 2002? Oh right, cheering on the Iraq War. Which has cost trillions of dollars and thousands of lives, and for what? For WHAT? Didn’t we supposedly invade Iraq for the “SECURITY” you now pooh-pooh, Mr. Brooks?

It’s almost hilarious that someone like David Brooks is cautioning the nation against becoming middle-aged, civilized and sedate. Of course he would create a false dichotomy that makes no sense at all. A quarter of the populace should remain uninsured for the sake of America’s VITALITY? A more “ragged” and unhealthy America is somehow more “vibrant.” But from what I can gather, he’s just tossing this out from his cozy insured cocoon. Other people, however, can and SHOULD struggle and make agonizing choices. This “vitality” is so essential it should come at their expense. But not SO essential that he should put his money where his mouth is and join them. I’m trying to parse this shit but my dendrites are having a breakdown. Maybe it’s just me, but aren’t people MORE productive when they’re NOT sick, and NOT worrying about getting fucked over by the entrenched interests squatting on top of the current health care system?

He also says this:

In the real world, there’s usually a trade-off. The unregulated market wants to direct capital to the productive and the young.

O RLY?

How does this guy have any credibility at all?!
Notes: 1 / 3 hours ago from sosexyoutofcontext

(via sosexyoutofcontext)
Notes: 3 / 5 hours ago from infectiousgroove (originally from aykyoto)

infectiousgroove:

aykyoto:

Vivid (via ♥Spice)
Notes: 207 / 7 hours ago from infectiousgroove (originally from kari-shma)

"What’s meant to be will always find a way"

- Trisha Yearwood (via kari-shma) (via infectiousgroove)
Notes: 9 / 9 hours ago from dailymeh
$dailymeh: Do the above equations hold the secret to consciousness? Probably not, though they may be our best guess. I’ll explain. People got excited when IBM announced that it had simulated a brain on the scale of a cat. I got excited. Henry Markram, who works on The Blue Brain Project, “the first comprehensive attempt to reverse-engineer the mammalian brain”, was not. He thinks it’s a fraud: These are point neurons (missing 99.999% of the brain; no branches; no detailed ion channels; the simplest possible equation you can imagine to simulate a neuron, totally trivial synapses; and using the STDP learning rule I discovered in this way is also is a joke). All these kinds of simulations are trivial and have been around for decades - simply called artificial neural network (ANN) simulations. (…) It is not even an innovation in simulation technology. You don’t need any special “C2 simulator”, this is just a hoax and a PR stunt. Most neural network simulators for parallel machines can can do this today. Nest, pNeuron, SPIKE, CSIM, etc, etc. all of them can do this! We could do the same simulation immediately, this very second by just loading up some network of points on such a machine, but it would just be a complete waste of time - and again, I would consider it shameful and unethical to call it a cat simulation. (…) This is light years away from a cat brain, not even close to an ants brain in complexity. Now, before we go on, I’d like to mention that I have absolutely no qualifications for running a science blog. I’m not a scientist, nor have I ever been one. I’m simply an enthusiastic layman. Unlike Adam, who also contributes to science tumbled, I’m not a PhD student in anything, much less regenerative medicine. If all of this meant that I was misleading people by passing on inaccurate sensationalism, I’m sorry. With that out of the way, we can take a look at what IBM’s simulation actually consists of. (And I’m still not a scientist, so these are still just my unnuanced thoughts.) The equations above are taken from The Cat Is Out of the Bag (pdf), the paper on the simulation. The paper explains the basics of how the simulation works: The basic algorithm of our cortical simulator C2 is that neurons are simulated in a clock-driven fashion whereas synapses are simulated in an event-driven fashion. For every neuron, at every simulation time step (say 1ms), we update the state of each neuron, and if the neuron fires, generate an event for each synapse that the neuron is post-synaptic and pre-synaptic to. For every synapse, when it receives a pre- or post-synaptic event, we update its state and, if necessary, the state of the post-synaptic neuron. Absolutely nothing remarkable here. It’s the most obvious way to do this, and while the specifics of how to scale this to hundreds of thousands of processors may be interesting from an engineering or computer science perspective, it doesn’t teach us anything about how the brain works. What’s interesting is the details of exactly how each neuron and synapse is updated at each time step, and how they’re connected. On the surface, at least, the way the simulation connects the neurons seems to be accurate as far as our current knowledge of how real brains are connected goes. (But again, someone who knows this stuff should look through the references.) The problem is in the simulation of each neuron and its synapses. According to the paper, each synapse is represented with a mere 16 bytes, to save space. Synapses are the spaces between neurons, through which a nerve cell sends neurotransmitters, the molecules that carry signals from one cell to another; the transmitters are then picked up by another cell and (potentially) send along the signal, called the action potential. Synapses are messy: sixteen bytes is probably not even close to capturing everything that goes on at each and every synapse. How accurate the model is depends on how much of the action is actually needed to produce cognition: maybe only some of the interactions at synapses are needed to create consciousness. But which interactions? Once you have synapses connecting neurons, you also need neurons actually processing the inputs. That’s what the equations above do. They simulate the signal processing that goes on in nerve cells; the simulation has four different types of neurons, and each is simulated by plugging in different parameters into the equations above. (At least that’s how I understand the paper’s appendix.) This is obviously a simplified model. Markram thinks it’s way too simple to actually amount to much. I could be very smug here and point to the fact that a simulation need not be accurate: a simulation is simply a model of some phenomenon, and even a very crude approximation of some phenomenon could be accurately called a simulation of that phenomenon (though not a very accurate simulation). I’m not going to; though it’s true that the simulation is a simulation, that’s not what all those journalists and bloggers (including me) wrote about, nor what people got excited about. The excitement was about the idea that this was a very good approximation of a real brain. This is what the leader of the IBM project has to say about just what is simulated: What aspects of the brain does the model include? The model reproduces a number of physiological and anatomical features of the mammalian brain. The key functional elements of the brain, neurons, and the connections between them, called synapses, are simulated using biologically derived models. The neuron models include such key functional features as input integration, spike generation and firing rate adaptation, while the simulated synapses reproduce time and voltage dependent dynamics of four major synaptic channel types found in cortex. Furthermore, the synapses are plastic, meaning that the strength of connections between neurons can change according to certain rules, which many neuroscientists believe is crucial to learning and memory formation. At an anatomical level, the model includes sections of cortex, a dense body of connected neurons where much of the brain’s high level processing occurs, as well as the thalamus, an important relay center that mediates communication to and from cortex. Much of the connectivity within the model follows a statistical map derived from the most detailed study to date of the circuitry within the cat cerebral cortex. So the idea that it’s “a cat’s brain” isn’t totally out there: the way the simulated neurons are connected is really based on statistical data derived from observations of real cats’ brains. The paper and the blog entry are very weak on just how the “input integration, spike generation and firing rate adaptation” were derived from actual data. Presumably, someone observed neurons in action and made the equations above to mimic what they saw, but I have no idea how they went from observation to model. That may be my fault: maybe looking at the references would find the procedure from observation to theoretical model perfectly well documented. To say that the simulation is completely unempirical would be unfair: We are able to observe activity in our model at many scales, ranging from global electrical activity levels, to activity levels in specific populations, to topographic activity dynamics to individual neuronal membrane potentials. In these measurements, we have observed the model reproduce activity in cortex measured by neuroscientists using corresponding techniques: electroencephalography, local field potential recordings, optical imaging with voltage sensitive dyes, and intracellular recordings. Specifically, we were able to deliver a stimulus to the model then watch as it propagated within and between different populations of neurons. We found that this propagation showed a spatiotemporal pattern remarkably similar to what has been observed in experiments with real brains. In other simulations, we also observed oscillations between active and quiet periods, as is often observed in the brain during sleep or quiet waking. So the model produces the same patterns as those found in real brains when subjected to similar stimuli. That would seem to indicate that the model, simplified though it may be, isn’t that bad of a guess. The advantage of such models seems to be that you get to record every neuron, every interaction that goes on, while you can only record a fraction of the data when observing a real brain. Of course, the data’s useless unless it matches what’s going on in a real brain. Markram’s Blue Brain project, meanwhile, isn’t content with equations that roughly model the real thing. Instead of modeling the neuron as an information processor by finding an equation that seems to produce the same outputs to the same inputs, they’re trying to model many of the biological processes that go on inside nerve cells. An interesting question is whether the neuron really is the fundamental unit of consciousness. Would any kind of signal processor that produces the same outputs to the same inputs as a neuron, connected in together with other similar processors in a way that models the brain, be conscious? Or do you need to simulate things on a smaller level still, perhaps the insides of cells, in order to produce conscious silicon? Or is there something magical about biological matter, some inherent property of carbon-based life that creates a consciousness that cannot be recreated by duplicating the signal-processing patterns in a computer?$
dailymeh:

Do the above equations hold the secret to consciousness? Probably not, though they may be our best guess. I’ll explain.

People got excited when IBM announced that it had simulated a brain on the scale of a cat. I got excited. Henry Markram, who works on The Blue Brain Project, “the first comprehensive attempt to reverse-engineer the mammalian brain”, was not. He thinks it’s a fraud:

These are point neurons (missing 99.999% of the brain; no branches; no detailed ion channels; the simplest possible equation you can imagine to simulate a neuron, totally trivial synapses; and using the STDP learning rule I discovered in this way is also is a joke).

All these kinds of simulations are trivial and have been around for decades - simply called artificial neural network (ANN) simulations. (…) It is not even an innovation in simulation technology. You don’t need any special “C2 simulator”, this is just a hoax and a PR stunt. Most neural network simulators for parallel machines can can do this today. Nest, pNeuron, SPIKE, CSIM, etc, etc. all of them can do this! We could do the same simulation immediately, this very second by just loading up some network of points on such a machine, but it would just be a complete waste of time - and again, I would consider it shameful and unethical to call it a cat simulation. (…) This is light years away from a cat brain, not even close to an ants brain in complexity.

Now, before we go on, I’d like to mention that I have absolutely no qualifications for running a science blog. I’m not a scientist, nor have I ever been one. I’m simply an enthusiastic layman. Unlike Adam, who also contributes to science tumbled, I’m not a PhD student in anything, much less regenerative medicine. If all of this meant that I was misleading people by passing on inaccurate sensationalism, I’m sorry.

With that out of the way, we can take a look at what IBM’s simulation actually consists of. (And I’m still not a scientist, so these are still just my unnuanced thoughts.) The equations above are taken from The Cat Is Out of the Bag (pdf), the paper on the simulation. The paper explains the basics of how the simulation works:

The basic algorithm of our cortical simulator C2 is that neurons are simulated in a clock-driven fashion whereas synapses are simulated in an event-driven fashion. For every neuron, at every simulation time step (say 1ms), we update the state of each neuron, and if the neuron fires, generate an event for each synapse that the neuron is post-synaptic and pre-synaptic to. For every synapse, when it receives a pre- or post-synaptic event, we update its state and, if necessary, the state of the post-synaptic neuron.

Absolutely nothing remarkable here. It’s the most obvious way to do this, and while the specifics of how to scale this to hundreds of thousands of processors may be interesting from an engineering or computer science perspective, it doesn’t teach us anything about how the brain works. What’s interesting is the details of exactly how each neuron and synapse is updated at each time step, and how they’re connected. On the surface, at least, the way the simulation connects the neurons seems to be accurate as far as our current knowledge of how real brains are connected goes. (But again, someone who knows this stuff should look through the references.) The problem is in the simulation of each neuron and its synapses.

According to the paper, each synapse is represented with a mere 16 bytes, to save space. Synapses are the spaces between neurons, through which a nerve cell sends neurotransmitters, the molecules that carry signals from one cell to another; the transmitters are then picked up by another cell and (potentially) send along the signal, called the action potential. Synapses are messy: sixteen bytes is probably not even close to capturing everything that goes on at each and every synapse. How accurate the model is depends on how much of the action is actually needed to produce cognition: maybe only some of the interactions at synapses are needed to create consciousness. But which interactions?

Once you have synapses connecting neurons, you also need neurons actually processing the inputs. That’s what the equations above do. They simulate the signal processing that goes on in nerve cells; the simulation has four different types of neurons, and each is simulated by plugging in different parameters into the equations above. (At least that’s how I understand the paper’s appendix.) This is obviously a simplified model. Markram thinks it’s way too simple to actually amount to much.

I could be very smug here and point to the fact that a simulation need not be accurate: a simulation is simply a model of some phenomenon, and even a very crude approximation of some phenomenon could be accurately called a simulation of that phenomenon (though not a very accurate simulation). I’m not going to; though it’s true that the simulation is a simulation, that’s not what all those journalists and bloggers (including me) wrote about, nor what people got excited about. The excitement was about the idea that this was a very good approximation of a real brain.

This is what the leader of the IBM project has to say about just what is simulated:

What aspects of the brain does the model include?

The model reproduces a number of physiological and anatomical features of the mammalian brain. The key functional elements of the brain, neurons, and the connections between them, called synapses, are simulated using biologically derived models. The neuron models include such key functional features as input integration, spike generation and firing rate adaptation, while the simulated synapses reproduce time and voltage dependent dynamics of four major synaptic channel types found in cortex. Furthermore, the synapses are plastic, meaning that the strength of connections between neurons can change according to certain rules, which many neuroscientists believe is crucial to learning and memory formation.

At an anatomical level, the model includes sections of cortex, a dense body of connected neurons where much of the brain’s high level processing occurs, as well as the thalamus, an important relay center that mediates communication to and from cortex. Much of the connectivity within the model follows a statistical map derived from the most detailed study to date of the circuitry within the cat cerebral cortex.

So the idea that it’s “a cat’s brain” isn’t totally out there: the way the simulated neurons are connected is really based on statistical data derived from observations of real cats’ brains. The paper and the blog entry are very weak on just how the “input integration, spike generation and firing rate adaptation” were derived from actual data. Presumably, someone observed neurons in action and made the equations above to mimic what they saw, but I have no idea how they went from observation to model. That may be my fault: maybe looking at the references would find the procedure from observation to theoretical model perfectly well documented.

To say that the simulation is completely unempirical would be unfair:

We are able to observe activity in our model at many scales, ranging from global electrical activity levels, to activity levels in specific populations, to topographic activity dynamics to individual neuronal membrane potentials. In these measurements, we have observed the model reproduce activity in cortex measured by neuroscientists using corresponding techniques: electroencephalography, local field potential recordings, optical imaging with voltage sensitive dyes, and intracellular recordings. Specifically, we were able to deliver a stimulus to the model then watch as it propagated within and between different populations of neurons. We found that this propagation showed a spatiotemporal pattern remarkably similar to what has been observed in experiments with real brains. In other simulations, we also observed oscillations between active and quiet periods, as is often observed in the brain during sleep or quiet waking.

So the model produces the same patterns as those found in real brains when subjected to similar stimuli. That would seem to indicate that the model, simplified though it may be, isn’t that bad of a guess. The advantage of such models seems to be that you get to record every neuron, every interaction that goes on, while you can only record a fraction of the data when observing a real brain. Of course, the data’s useless unless it matches what’s going on in a real brain.

Markram’s Blue Brain project, meanwhile, isn’t content with equations that roughly model the real thing. Instead of modeling the neuron as an information processor by finding an equation that seems to produce the same outputs to the same inputs, they’re trying to model many of the biological processes that go on inside nerve cells. An interesting question is whether the neuron really is the fundamental unit of consciousness. Would any kind of signal processor that produces the same outputs to the same inputs as a neuron, connected in together with other similar processors in a way that models the brain, be conscious? Or do you need to simulate things on a smaller level still, perhaps the insides of cells, in order to produce conscious silicon? Or is there something magical about biological matter, some inherent property of carbon-based life that creates a consciousness that cannot be recreated by duplicating the signal-processing patterns in a computer?
Notes: 18 / 11 hours ago from thoughtsdetained

"there are two kinds of people in this world: those who believe there are two kinds of people in this world and those who are smart enough to know better."

- Tim Robbins, Still Life of a Wood Pecker (via thoughtsdetained)
Notes: 4 / 13 hours ago from smut-to-go

"She had been getting ready for her great journey to the horizons in search of people; it was important to all the world that she should find them and they find her. But she had been whipped like a cur dog, and run off down a back road after things. It was all according to the way you see things. Some people could look at a mud-puddle and see an ocean with ships. But Nanny belonged to that other kind that loved to deal in scraps. Here Nanny had taken the biggest thing God ever made, the horizon - for no matter how far a person can go the horizon is still way behind you - and pinched it in to such a little bit of a thing that she could tie it about her granddaughter’s neck tight enough to choke her…She had found a jewel down inside herself and she had wanted to walk where people could see her and gleam it around. But she had been set in the market-place to sell. Been set for still-bait. When God had made The Man, he made him out of stuff that sung all the time and glittered all over. Then after that some angels got jealous and chopped him into millions of pieces, but still he glittered and hummed. So they beat him down to nothing but sparks but each little spark had a shine and a song. So they covered each one over with mud. And the lonesomeness in the sparks made them hunt for one another, but the mud is deaf and dumb. Like all the other tumbling mud-balls, Janie had tried to show her shine."

- Their Eyes Were Watching God [emphasis is mine] (via smut-to-go)
Notes: 1 / 15 hours ago from jmarieb

(via jmarieb)
Notes: 772 / 17 hours ago from scout (originally from thedailywhat)

scout:

byronic / thedailywhat:

The entire cast of The Muppet Show perform the “Bohemian Rhapsody” cover to end all “Bohemian Rhapsody” covers.

[via.]

oh my godddd. laughing out loud in the library.
Notes: 34 / 19 hours ago from libraryland (originally from uncertaintimes)

libraryland:

uncertaintimes:

the music of sound

Following

Tumblr