And this can't be a one-time deal. In my field anyway (neuroscience & psychology) there is a legitimate need for revisions every few years as the science is progressing quickly.
If the states are serious about this, they should pay good salaries to teams of professors, editors, graphic designers, etc... to create high quality material that they then provide by open access. The current open access textbooks available in my field (neuroscience & psychology) are horrible. This would save the students significant money, but would cost the states more. Otherwise, this is just a combination of talk and trying to get people to do work for free.
When on the drugs the Lempel-Ziv (LZ) complexity score was higher, but the paper also states: "Since the value of LZc (also LZs) for a binary sequence of fixed length is maximal if the sequence is entirely random, the normalized values indicate the level of signal diversity on a scale from 0 to 1." In other words, higher LZ means more random. However, "true" complexity does not increase monotonically with randomness. In fact, they have an inverted U relationship, where past a certain point, complexity _decreases_ with increasing randomness. For example, see Figure 2 on this page: http://www.cs.brandeis.edu/~pa... I would argue that consciousness if far more likely to relate to this type of "true" complexity rather than the uncompressability expressed by LZ.
The comparison of "deep learning that needs tons of examples" vs "Bayesian programming that can learn from a few examples" is a false dichotomy. It all depends on how much structure you assume a priori versus how much structure you learn from the data.
Typical neural net (deep learning) examples start with no structure and thus require lots of examples. Typical Bayesian net examples start with a lot of structure and thus require only a few examples.
On the other hand, if you start with a highly pretrained net like Inception-v3, then your deep learning cat expert may not need as many examples to generalize. And if your Bayesian programming model starts out with very general, very simple "building blocks" then it may need a lot of examples to extract the predictable structure.
A main difference is that if you want to start with a lot of structure built in, you will probably have to pretrain for the neural net, whereas you can "hand code" the knowledge in your Bayes net. And the structure in the Bayes net may be a lot more transparent and easily interpretable than in the neural net. On the other hand, you had better hope you picked the right structure to begin with or else you will be reasoning over possibilities that are all very wrong! Knowing that an image is 50 times more likely to be a cat than a dog is not very helpful if it is actually a penguin.
To understand Trump, you must appreciate that he is not a liar, he is a bullshitter: "Bullshit can be neither true nor false; hence, the bullshitter is someone whose principal aim—when uttering or publishing bullshit—is to impress the listener and the reader with words that communicate an impression that something is being or has been done, words that are neither true nor false, and so obscure the facts of the matter being discussed. In contrast, the liar must know the truth of the matter under discussion, in order to better conceal it from the listener or the reader being deceived with a lie; while the bullshitter’s sole concern is personal advancement and advantage to their own agenda." (https://en.wikipedia.org/wiki/On_Bullshit) The truth value of Trump's statements are irrelevant to Trump and, apparently, his supporters. Once you accept this, you realize that debating about Trump on the basis of facts is pointless. This is the challenge of dealing with Trump.
This study is the latest in a long line of studies which show that we have poor insight into the causes of our choices. We think we know why we have decided the way we have, but experiment after experiment has shown that this is often not entirely true. This study is showing that even events which occur _after_ the moment when we think we have decided can influence our decisions without our being aware of this fact.
Our explanations of why we have decided the way we have are stories that we construct. Whether these stories are constructed before or after the decision is made, they are still stories that do not fully account for the actual causes of the choices.
Our sense of free will is grounded in the idea that we know why we make the choices we do. This experiment contributes to a large body of literature which demonstrates that this is simply not true.
Sorry, this article is not an intelligent take on artificial intelligence.
First, the Chinese Room argument applies equally to whatever new algorithm Jeff Hawkins comes up with as it does to the English speaking man plus book algorithm. The Chinese Room argument applies equally to ANY algorithmic approach to intelligence. Thus, his call for a new definition of intelligence will do nothing to disarm the Chinese Room argument.
Second, the Chinese Room argument has already been thoroughly addressed by Daniel Dennett and others. If you really care, go read Searle's The Mystery of Consciousness and Dennett's Consciousness Explained, and then go from there.
Third, in the mean time, researchers will continue on their merry way toward developing AI. Guess what? It is harder than Minsky and others ever imagined, but none the less, we are making progress. And guess what else? When we have "truly" intelligent systems, there won't be any magic in there. Just boring, dumb algorithms. Our robot overlords will be Chinese Rooms.
We all pose, because the "self" is nothing but a pose. To think that there is a way to be a living human and not pose is naive, and to berate others for posing is just silly.
Actually, what Apple does is to create computing _appliances_ with intuitive interfaces. Hype not withstanding, the reason why Apples devices are successful is because they ARE easier and simpler to use. All of the control that Apple exerts over the ecosystems for their devices has one primary aim: simplicity and consistency. This has been going on since the elaborate "Macintosh Human Interface Guidelines" for the original MacOS. This is why their mice have one button. This is why they resist third-party clones. This is why they have a single iTunes store. Etc, etc... By standardizing and making many choices for the user, they allow the user to focus on their non-technical goals: writing a document, listening to music, etc... For those of us who LIKE to have complete control over the technical details of our devices, and like to spend time making choices about technical matters, this can be exceedingly frustrating. I wish I could modify the OS on my wrist watch as well. I'm a freak. So are you. We aren't Apple's target market. Get over it.
Any chance they have fixed quoted-line prefixing for replies and added "Paste as Quotation" to Outlook? That alone would be worth an upgrade. Yes, I know about the "Prefix each line of the original message" option. And no, I don't care whether you like inline replies or not, just make it reasonably easy to use them, and give me the choice. Is this really so hard?
I'm not surprised at all. Having diligently tried to use IE8 for months, I can confidently say it is a horrible experience. Much worse than IE6 ever was. It hangs on a regular basis - not just one tab, but the whole progam. New tabs can take a long time to come up. It slowly eats more and more memory. I've experienced bizarre bugs, for example I load a page that renders incorrectly, I click through to another page, and then hit the 'back' button, and the first page now renders correctly. Etc... Microsoft is driving down their own market share by providing a shitty product. We're not talking bells and whistles here, just basic functionality.
Anyway, you're missing the point. The savings are realized by individual employers and consumers at the critical point of decision making about how to spend their money. The costs are realized indirectly and in a distributed fashion across society by higher government expenses leading to higher taxes. As a result, the employers and consumers are making decisions that are NOT in their own best interests, because the meager savings are obvious and immediate while the high costs are indirect and delayed. I'm not arguing that this situation is rational, in fact, I'm arguing the opposite, people need to wake up to the true costs and act accordingly.
I repeatedly see people discussing the costs illegal immigrant incur on society, particularly since they typically don't have car insurance or medical insurance. What I don't see is a discussion of the demand driving them to be here: 1. Employers, who are US citizens, seek them out because they are cheap labor and can be abused without repercussions, since they are afraid to report the abuse to authorities. 2. Consumers, who are US citizens, choose to purchase the goods and services created by their labor, because they are cheaper than the competing products.
Did you buy a newly built home in the last 10 years? Did you visit the construction site? Can you honestly say you were not aware that many of the day laborers working on the site where probably illegal? Did you do anything about it?
Have you bought berries on sale at the grocery store brought in from California? Can you honestly say you weren't aware that they were probably picked by illegal immigrants?
If we really want to quell the flow of illegal immigrants across the border, we need to address the demand for their services. Take an active role in NOT buying goods and services provided by them, and the demand will go down. Yes, its going to cost you money. Lots of money. That's the reality - we will all pay more for the goods and services currently provided by illegal immigrants if they are instead provided by legal immigrants and US citizens. Let's talk about that for a while.
Or maybe the obese rates are those that had no self-control to start out with. If that's the case, the severe obesity might simply be a visible indicator of a very real character flaw. (Although I have serious questions about the meaning of "moral failure", if brain chemistry determines a person's actions.)
First, the obese rats were the ones in the "eat sugary fatty food" condition, they were not self-selected. Second, if brain chemistry doesn't determine one's actions, what does? Are you a dualist or something?
'Why should you ever, with all this parallel hardware, ever be waiting for your computer?'
In my experience, the vast majority of waiting that I do is because of hard disk I/O, not computation. I suspect that focusing effort on the intelligent and wide-spread use of SSDs and other faster-than-hard-drive media will do more to minimize my wait time.
If you do your data analysis in R and write your paper in latex, then you can use sweave to create a single file with the R code embedded in the latex. When you process the latex file, the R code is run, generating your stats and figures on the fly for the resulting document. If this file (plus any data files it reads, and any non-standard code it calls) is posted as "supplementary material" along with the PDF journal article, then there is no question about the software or analysis that led to the figures in the paper. Of course, the data itself is still open to question.
I hope they work on stability, memory usage, and the pre-rendering speed! I've found IE8 to be less stable and use more memory than recent past versions of IE. And just getting a new blank tab to come up often involves a fair amount of thumb-twiddling. And despite whatever usage of different processes for different tabs they claim to be employing, I find that the entire browser usually hangs and crashes when there is a problem with a page. Rendering speed and font readability are the least of their problems!
One state is openly hostile to the idea of freedom. The other is merely indifferent.
Uh, you seem to be confused. The ones fighting for freedom here are those trying to stop these bigoted, constraining laws.
Attempting to use the legal system to take away the rights of a minority is the worst kind of violence. I have no sympathy for people who are doing that or their apologists.
Its a fact that FICA and Medicare often don't have to be paid for international students. This is federal law, so it's not surprising that more than one university describes the same factual situation that applies across the country. This is not under the control of the universities.
Note that, due to the various issues with visas, paperwork, etc., international students often struggle to find employment, and so its not unreasonable for universities to advocate on their behalf. Universities also typically have an entire career services group that helps all the American students, so they are hardly neglected.
It takes a fair amount of cynicism (I know, I know, its/.) to spin a university's attempt to support their international students into an attack on their American students.
If you really care about this issue, you should just aggitate against the tax law itself.
But as a consumer I don't want my maximum Mbps download rate to be limited based on the assumption that I will be using the full bandwidth all of the time. Neither I nor the vast majority of users behave in this way. But when I do want to download something, I'd like it to be fast. So I'd much rather pay for a shared pipe with some reasonable rules about how it is shared so that I don't get screwed by the one guy next door who is always downloading as much as possible. I would much rather have a higher maximum Mbps coupled with a reasonable limit on total consumption. I'll pay less because I'm willing to share.
If you really want a high bandwidth pipe all to yourself, then you should pay more.
Of course, the options should be stated in a much clearer and more honest way by the providers than they typically are currently. For example, a particular package could include: 1Mbps of guaranteed bandwidth, up to 40 Mbps of maximum bandwidth contingent on what the neighbors are up to, and 100Gb per month cap.
The example video shows discrimination of 4 available discrete actions. The eventual goal would presumably be to discriminate tens, hundreds, or thousands of actions, if not smoothly varying parameters of action.
There are two main ways to go about this: 1. Train the algorithms processing the brain signals. 2. Train the brain signals.
The best approach is probably to do both in concert using real-time feedback to the user about how the algorithm is currently interpreting the signal. The user can then learn (explicitly or implicitly) what mental processes lead to what action outcomes. In this way the user will learn to control the system in much the same way we learn and constantly relearn how to control our own limbs.
A significant issue will be how good the signal is that can be obtained with EEG and NIRS. Hopefully EEG alone is enough, because while NIRS takes some significant hardware, you could wear an EEG cap around all day pretty easily if you really had to.
"the phantom limb of her now-paralyzed left arm could... 'scratch an itch on her head, with an actual sense of relief.'"
This seems consistent with an evolving view of perception as a Bayesian process that incorporates both the current sensory evidence and our prior knowledge of what has been perceived in similar situations in the past. This is coupled with the additional fact that information about motor movements is directly provided to sensory parts of the brain from motor parts of the brain without needing to be perceived through external sensory systems. When a limb is missing due to amputation, or the sensory input from it is missing due to paralysis, the brain is still sending motor information (despite the fact the motor signals aren't actually able to get to the missing/paralyzed limb) and it still has all of the prior knowledge about what has previously been experienced when such a motor action is taken. As a result, when a phantom limb is used to scratch an itch, the perceptual system 'knows' the scratching action is being taken and 'knows' that scratching has previously relieved itching, and despite the fact that no actual sensory info to this extent is coming from the head, the Bayesian posterior still makes 'itch relief' be the most likely outcome, and thus, apparently, that is the experienced outcome.
The main point to take from this is that perception, while certainly related to what is actually going on in the external world, is far far FAR from a 'pure' reflection of it.
"The Javascript Trap" seems reasonable to me. Stallman wants the ability to control the software that runs on HIS computer. He is simply calling for a convenient mechanism to:
1. Allow me to set a policy to only run code when I have access to the source. 2. Allow me to swap out any code that will run on my computer with another chunk of code of my choosing.
Neither of these things seems outlandish or unreasonable to me. And both seem like things that could be implemented fairly easily in a browser. If you don't care about this, then carry on as you are. If you do care about controlling and being able to view the code that runs on your computer, then such a mechanism would be quite handy.
The form of free will in the TFA is summarized in the line: "the choice an experimenter makes is not a function of the past" The article argues that if we have this form of free will, then elementary particles must have it as well.
I, for one, am satisfied with a much different form of free will: "the choice an experimenter makes cannot be predicted in advanced by anyone (including the experimenter him/herself)" For this, we need only acknowledge that to predict what decision an experimenter would make at a given time, we would need to know in advance the experimenter's exact brain state at the time the decision is made. And in order to know that, we would need to know the exact state at some earlier time, and then run a model of the brain and the environment with which it interacts up to the point of the decision. How detailed would the model need to be? How detailed would our knowledge of the earlier state and the environment need to be? To make the prediction with certainty, they would need to be exact. How could the experimenter and his environment be modelled exactly? There is only one way: the actual experimenter and environment (or a perfect replica), occuring in real time. Thus I submit to you that the decision is unknowable by anyone (including the experimenter) until it actually occurs.
This form of free will is perfectly compatible with a fully deterministic universe, while still being fully satisfying (to me anyway).
Slashdot Mirror
And this can't be a one-time deal. In my field anyway (neuroscience & psychology) there is a legitimate need for revisions every few years as the science is progressing quickly.
If the states are serious about this, they should pay good salaries to teams of professors, editors, graphic designers, etc... to create high quality material that they then provide by open access. The current open access textbooks available in my field (neuroscience & psychology) are horrible. This would save the students significant money, but would cost the states more. Otherwise, this is just a combination of talk and trying to get people to do work for free.
When on the drugs the Lempel-Ziv (LZ) complexity score was higher, but the paper also states:
"Since the value of LZc (also LZs) for a binary sequence of fixed length is maximal if the sequence is entirely random, the normalized values indicate the level of signal diversity on a scale from 0 to 1."
In other words, higher LZ means more random. However, "true" complexity does not increase monotonically with randomness. In fact, they have an inverted U relationship, where past a certain point, complexity _decreases_ with increasing randomness. For example, see Figure 2 on this page:
http://www.cs.brandeis.edu/~pa...
I would argue that consciousness if far more likely to relate to this type of "true" complexity rather than the uncompressability expressed by LZ.
The comparison of "deep learning that needs tons of examples" vs "Bayesian programming that can learn from a few examples" is a false dichotomy. It all depends on how much structure you assume a priori versus how much structure you learn from the data.
Typical neural net (deep learning) examples start with no structure and thus require lots of examples. Typical Bayesian net examples start with a lot of structure and thus require only a few examples.
On the other hand, if you start with a highly pretrained net like Inception-v3, then your deep learning cat expert may not need as many examples to generalize. And if your Bayesian programming model starts out with very general, very simple "building blocks" then it may need a lot of examples to extract the predictable structure.
A main difference is that if you want to start with a lot of structure built in, you will probably have to pretrain for the neural net, whereas you can "hand code" the knowledge in your Bayes net. And the structure in the Bayes net may be a lot more transparent and easily interpretable than in the neural net. On the other hand, you had better hope you picked the right structure to begin with or else you will be reasoning over possibilities that are all very wrong! Knowing that an image is 50 times more likely to be a cat than a dog is not very helpful if it is actually a penguin.
To understand Trump, you must appreciate that he is not a liar, he is a bullshitter:
"Bullshit can be neither true nor false; hence, the bullshitter is someone whose principal aim—when uttering or publishing bullshit—is to impress the listener and the reader with words that communicate an impression that something is being or has been done, words that are neither true nor false, and so obscure the facts of the matter being discussed. In contrast, the liar must know the truth of the matter under discussion, in order to better conceal it from the listener or the reader being deceived with a lie; while the bullshitter’s sole concern is personal advancement and advantage to their own agenda." (https://en.wikipedia.org/wiki/On_Bullshit)
The truth value of Trump's statements are irrelevant to Trump and, apparently, his supporters. Once you accept this, you realize that debating about Trump on the basis of facts is pointless. This is the challenge of dealing with Trump.
This study is the latest in a long line of studies which show that we have poor insight into the causes of our choices. We think we know why we have decided the way we have, but experiment after experiment has shown that this is often not entirely true. This study is showing that even events which occur _after_ the moment when we think we have decided can influence our decisions without our being aware of this fact.
Our explanations of why we have decided the way we have are stories that we construct. Whether these stories are constructed before or after the decision is made, they are still stories that do not fully account for the actual causes of the choices.
Our sense of free will is grounded in the idea that we know why we make the choices we do. This experiment contributes to a large body of literature which demonstrates that this is simply not true.
Sorry, this article is not an intelligent take on artificial intelligence.
First, the Chinese Room argument applies equally to whatever new algorithm Jeff Hawkins comes up with as it does to the English speaking man plus book algorithm. The Chinese Room argument applies equally to ANY algorithmic approach to intelligence. Thus, his call for a new definition of intelligence will do nothing to disarm the Chinese Room argument.
Second, the Chinese Room argument has already been thoroughly addressed by Daniel Dennett and others. If you really care, go read Searle's The Mystery of Consciousness and Dennett's Consciousness Explained, and then go from there.
Third, in the mean time, researchers will continue on their merry way toward developing AI. Guess what? It is harder than Minsky and others ever imagined, but none the less, we are making progress. And guess what else? When we have "truly" intelligent systems, there won't be any magic in there. Just boring, dumb algorithms. Our robot overlords will be Chinese Rooms.
We all pose for society.
We all pose, because the "self" is nothing but a pose. To think that there is a way to be a living human and not pose is naive, and to berate others for posing is just silly.
>>And the undeniable fact is that virtually all racists have joined the Republican party
All 30 of them left in America?
You can't possibly be serious. You are, indeed, an anonymous coward.
Wow.
Actually, what Apple does is to create computing _appliances_ with intuitive interfaces. Hype not withstanding, the reason why Apples devices are successful is because they ARE easier and simpler to use. All of the control that Apple exerts over the ecosystems for their devices has one primary aim: simplicity and consistency. This has been going on since the elaborate "Macintosh Human Interface Guidelines" for the original MacOS. This is why their mice have one button. This is why they resist third-party clones. This is why they have a single iTunes store. Etc, etc... By standardizing and making many choices for the user, they allow the user to focus on their non-technical goals: writing a document, listening to music, etc... For those of us who LIKE to have complete control over the technical details of our devices, and like to spend time making choices about technical matters, this can be exceedingly frustrating. I wish I could modify the OS on my wrist watch as well. I'm a freak. So are you. We aren't Apple's target market. Get over it.
Any chance they have fixed quoted-line prefixing for replies and added "Paste as Quotation" to Outlook? That alone would be worth an upgrade. Yes, I know about the "Prefix each line of the original message" option. And no, I don't care whether you like inline replies or not, just make it reasonably easy to use them, and give me the choice. Is this really so hard?
I'm not surprised at all. Having diligently tried to use IE8 for months, I can confidently say it is a horrible experience. Much worse than IE6 ever was. It hangs on a regular basis - not just one tab, but the whole progam. New tabs can take a long time to come up. It slowly eats more and more memory. I've experienced bizarre bugs, for example I load a page that renders incorrectly, I click through to another page, and then hit the 'back' button, and the first page now renders correctly. Etc... Microsoft is driving down their own market share by providing a shitty product. We're not talking bells and whistles here, just basic functionality.
you fucking slashtard!!!!!!!!!!
How insightful, or is that inciteful?
Anyway, you're missing the point. The savings are realized by individual employers and consumers at the critical point of decision making about how to spend their money. The costs are realized indirectly and in a distributed fashion across society by higher government expenses leading to higher taxes. As a result, the employers and consumers are making decisions that are NOT in their own best interests, because the meager savings are obvious and immediate while the high costs are indirect and delayed. I'm not arguing that this situation is rational, in fact, I'm arguing the opposite, people need to wake up to the true costs and act accordingly.
I repeatedly see people discussing the costs illegal immigrant incur on society, particularly since they typically don't have car insurance or medical insurance. What I don't see is a discussion of the demand driving them to be here:
1. Employers, who are US citizens, seek them out because they are cheap labor and can be abused without repercussions, since they are afraid to report the abuse to authorities.
2. Consumers, who are US citizens, choose to purchase the goods and services created by their labor, because they are cheaper than the competing products.
Did you buy a newly built home in the last 10 years? Did you visit the construction site? Can you honestly say you were not aware that many of the day laborers working on the site where probably illegal? Did you do anything about it?
Have you bought berries on sale at the grocery store brought in from California? Can you honestly say you weren't aware that they were probably picked by illegal immigrants?
If we really want to quell the flow of illegal immigrants across the border, we need to address the demand for their services. Take an active role in NOT buying goods and services provided by them, and the demand will go down. Yes, its going to cost you money. Lots of money. That's the reality - we will all pay more for the goods and services currently provided by illegal immigrants if they are instead provided by legal immigrants and US citizens. Let's talk about that for a while.
Or maybe the obese rates are those that had no self-control to start out with. If that's the case, the severe obesity might simply be a visible indicator of a very real character flaw. (Although I have serious questions about the meaning of "moral failure", if brain chemistry determines a person's actions.)
First, the obese rats were the ones in the "eat sugary fatty food" condition, they were not self-selected. Second, if brain chemistry doesn't determine one's actions, what does? Are you a dualist or something?
'Why should you ever, with all this parallel hardware, ever be waiting for your computer?'
In my experience, the vast majority of waiting that I do is because of hard disk I/O, not computation. I suspect that focusing effort on the intelligent and wide-spread use of SSDs and other faster-than-hard-drive media will do more to minimize my wait time.
If you do your data analysis in R and write your paper in latex, then you can use sweave to create a single file with the R code embedded in the latex. When you process the latex file, the R code is run, generating your stats and figures on the fly for the resulting document. If this file (plus any data files it reads, and any non-standard code it calls) is posted as "supplementary material" along with the PDF journal article, then there is no question about the software or analysis that led to the figures in the paper. Of course, the data itself is still open to question.
I hope they work on stability, memory usage, and the pre-rendering speed! I've found IE8 to be less stable and use more memory than recent past versions of IE. And just getting a new blank tab to come up often involves a fair amount of thumb-twiddling. And despite whatever usage of different processes for different tabs they claim to be employing, I find that the entire browser usually hangs and crashes when there is a problem with a page. Rendering speed and font readability are the least of their problems!
One state is openly hostile to the idea of freedom. The other is merely indifferent.
Uh, you seem to be confused. The ones fighting for freedom here are those trying to stop these bigoted, constraining laws. Attempting to use the legal system to take away the rights of a minority is the worst kind of violence. I have no sympathy for people who are doing that or their apologists.
Wow, great spin job!
Its a fact that FICA and Medicare often don't have to be paid for international students. This is federal law, so it's not surprising that more than one university describes the same factual situation that applies across the country. This is not under the control of the universities.
Note that, due to the various issues with visas, paperwork, etc., international students often struggle to find employment, and so its not unreasonable for universities to advocate on their behalf. Universities also typically have an entire career services group that helps all the American students, so they are hardly neglected.
It takes a fair amount of cynicism (I know, I know, its /.) to spin a university's attempt to support their international students into an attack on their American students.
If you really care about this issue, you should just aggitate against the tax law itself.
But as a consumer I don't want my maximum Mbps download rate to be limited based on the assumption that I will be using the full bandwidth all of the time. Neither I nor the vast majority of users behave in this way. But when I do want to download something, I'd like it to be fast. So I'd much rather pay for a shared pipe with some reasonable rules about how it is shared so that I don't get screwed by the one guy next door who is always downloading as much as possible. I would much rather have a higher maximum Mbps coupled with a reasonable limit on total consumption. I'll pay less because I'm willing to share.
If you really want a high bandwidth pipe all to yourself, then you should pay more.
Of course, the options should be stated in a much clearer and more honest way by the providers than they typically are currently. For example, a particular package could include: 1Mbps of guaranteed bandwidth, up to 40 Mbps of maximum bandwidth contingent on what the neighbors are up to, and 100Gb per month cap.
The example video shows discrimination of 4 available discrete actions. The eventual goal would presumably be to discriminate tens, hundreds, or thousands of actions, if not smoothly varying parameters of action.
There are two main ways to go about this:
1. Train the algorithms processing the brain signals.
2. Train the brain signals.
The best approach is probably to do both in concert using real-time feedback to the user about how the algorithm is currently interpreting the signal. The user can then learn (explicitly or implicitly) what mental processes lead to what action outcomes. In this way the user will learn to control the system in much the same way we learn and constantly relearn how to control our own limbs.
A significant issue will be how good the signal is that can be obtained with EEG and NIRS. Hopefully EEG alone is enough, because while NIRS takes some significant hardware, you could wear an EEG cap around all day pretty easily if you really had to.
"the phantom limb of her now-paralyzed left arm could... 'scratch an itch on her head, with an actual sense of relief.'"
This seems consistent with an evolving view of perception as a Bayesian process that incorporates both the current sensory evidence and our prior knowledge of what has been perceived in similar situations in the past. This is coupled with the additional fact that information about motor movements is directly provided to sensory parts of the brain from motor parts of the brain without needing to be perceived through external sensory systems. When a limb is missing due to amputation, or the sensory input from it is missing due to paralysis, the brain is still sending motor information (despite the fact the motor signals aren't actually able to get to the missing/paralyzed limb) and it still has all of the prior knowledge about what has previously been experienced when such a motor action is taken. As a result, when a phantom limb is used to scratch an itch, the perceptual system 'knows' the scratching action is being taken and 'knows' that scratching has previously relieved itching, and despite the fact that no actual sensory info to this extent is coming from the head, the Bayesian posterior still makes 'itch relief' be the most likely outcome, and thus, apparently, that is the experienced outcome.
The main point to take from this is that perception, while certainly related to what is actually going on in the external world, is far far FAR from a 'pure' reflection of it.
"The Javascript Trap" seems reasonable to me. Stallman wants the ability to control the software that runs on HIS computer. He is simply calling for a convenient mechanism to:
1. Allow me to set a policy to only run code when I have access to the source.
2. Allow me to swap out any code that will run on my computer with another chunk of code of my choosing.
Neither of these things seems outlandish or unreasonable to me. And both seem like things that could be implemented fairly easily in a browser. If you don't care about this, then carry on as you are. If you do care about controlling and being able to view the code that runs on your computer, then such a mechanism would be quite handy.
The form of free will in the TFA is summarized in the line:
"the choice an experimenter makes is not a function of the past"
The article argues that if we have this form of free will, then elementary particles must have it as well.
I, for one, am satisfied with a much different form of free will:
"the choice an experimenter makes cannot be predicted in advanced by anyone (including the experimenter him/herself)"
For this, we need only acknowledge that to predict what decision an experimenter would make at a given time, we would need to know in advance the experimenter's exact brain state at the time the decision is made. And in order to know that, we would need to know the exact state at some earlier time, and then run a model of the brain and the environment with which it interacts up to the point of the decision. How detailed would the model need to be? How detailed would our knowledge of the earlier state and the environment need to be? To make the prediction with certainty, they would need to be exact. How could the experimenter and his environment be modelled exactly? There is only one way: the actual experimenter and environment (or a perfect replica), occuring in real time. Thus I submit to you that the decision is unknowable by anyone (including the experimenter) until it actually occurs.
This form of free will is perfectly compatible with a fully deterministic universe, while still being fully satisfying (to me anyway).