When is the last time you negotiated prices at the grocery store?
Every time you buy something at the grocery store, you are negotiating that item's price. If you refuse to buy an item because it's overpriced, the company that manufactures that item has the choice to either lower the price, try to market to a demographic that's willing to pay a higher price, or to attempt to keep selling an item at a price point at which the item won't sell.
By buying or not buying an item, you help determine demand for that item, thereby, setting its price. Basic stuff. Same set of principles applies to the labor market.
A few things about the reddit situation:
- reddit's staff will become less talented over time, as long as this policy persists, until they find some talent-price equilibrium that reflects reddit's salary policy.
- the CEO that implemented this policy will not suffer financially for doing so. However, it is likely that the talent of reddit's staff will decrease over time, and may result in her premature dismissal, possibly with an overly generous golden parachute.
- this salary policy has only a tenuous connection, if any, to the issue of women and men receiving an equal salary for equal labor. If reddit's CEO is not an absolute fool, the gender equality explanation is a rationalization intended to make an ill-conceived cost-cutting measure more palatable. If reddit's CEO is sincere in her explanation for this move, the board should remove her (and other companies should avoid her), because she does not understand certain fundamental labor market principles.
Here's another prediction: reddit will not save as much money as they think they will, because their HR department will be forced to dedicate more time to efforts to acquire adequate talent.
Have you met some of the MBAs who teach business courses? For a shock, try asking a few of them some fundamental stats questions that a person who has taken some grad-level stats courses (a prerequisite for many scientific/quantitative fields) should be able to answer. I can tell you about MBA profs who use statistical analysis allegedly on a regular basis without knowing the term "R^2".
Ballmer's probably a step up from quite a few people career academics in the business field.
Biz 101: How to ingratiate yourself with the right people Biz 324: Managing technical staff for the non-technical (focus on developers) Biz 412: Diversifying your product offerings, and what to do when you fail at it
The hard part is motivation. The problem with school is the knowledge looks "fake" to the majority of young schoolchildren, because it has no apparent connection what adults do. Small children are very quick to get excited about mimicking "real" things -- things they observe adults doing.
Agree with a notion that extends what you said there. Children have less ideas/concepts/mental schemas on which to draw, so much of what is presented to them in school appears arbitrary. One of the paths towards more effective education is to associate new knowledge with concepts that are present in a student's mind. This is pretty basic stuff that William James was saying in the 19th century, and which has a more modern instantiation in our understanding of artificial and biological neural networks. In a nutshell, when teachers present ideas, if it seems that the idea appeared out of a vacuum, or is arbitrary, the student will either be skeptical about the idea, or will have a hard time integrating it (i.e. may have to take the idea on faith and rotely memorize it, and retention will be difficult due to a lack of association with other concepts). The process of relating new ideas to ideas that have already been integrated by a person is crucial to the process of learning.
I honestly don't see how people are making it. I think the best teachers now go to private schools or colleges, and many (but not all, mind you) of the ones who remain are the ones who just aren't very good.
Correct me if I'm wrong, but I've read that many teachers who move to private schools from public institutions take a pay cut to do so, and they do it because they find the environment more comfortable and more conducive to doing their job. The students tend to be more classwork-oriented, and they have less disciplinary issues, so although there's less financial incentive to work in private schools (including no prospect of a nice pension upon retirement), the experience itself is more gratifying. Teachers in some urban areas, such as Chicago, for example, can make six figure salaries, along with some very nice perks; I'm not sure how that has worked out for them in terms of educational outcomes.
It seems that every aspect of the education experience feeds into every other aspect. Indifferent parents (as you've pointed out) produce indifferent or downright hostile kids, who, in turn, make many of the best teachers run for the hills, which leaves the parents and kids to deal with the less well-equipped teachers, and this process propagates through the generations. Most of the fix to education should will probably come with a cultural shift, rather than by throwing more technology into the classroom. People were learning trigonometry and calculus just fine before they had computers in the classroom, and there's no reason they can't do so now. Use computers when necessary (i.e. working with complex regression equations in a Stats course), and leave them out of the classroom when they're simply expedient or a hindrance. Things such as kindles, which present a great variety of books to students, can be extremely useful. Things like iPads are attention hogs, and wire young brains in ways that may not help them in their future endeavors. Use technology as appropriate, but don't pretend that it will be a panacea.
The unsupervised algorithm discussed in the article seems to code some sort of visual input and, I'd infer, to perform clustering, which permits it to assign labels (i.e. let's say, 'tree', 'human', etc.) to objects it has encountered. It can use this schema which it has constructed to assign objects it hasn't seen before to a cluster - that is, it labels novel inputs in accordance with its schema. Thus, the algorithm 'recognizes' classes of objects. I'd imagine if you granularize the detail level of detail to which the algorithm pays attention, it can recognize (categories of one object)/particular instances of a general object category/a particular object.
Sitting on a street, a "bicycle" is an object because it is most like to be operated on as a unit. But to a bicycle mechanic, a bicycle is a collection of objects, such as a frame, a seat.. and so on because they need to decompose the "bicycle" construct to do their job. To somebody on an assembly line putting together bicycle seats, a seat is (at least initially) several different objects.
That goes way beyond the task of 'classifying objects', which is what the algorithm is intended to do. It's like expecting a mechanic to drive your vehicle for you once he has repaired it.
So, truly unsupervised algorithms cannot do useful recognition - that is, classify objects the same way people do. (A robot that could experiment with its environment and learn to use "objects" could come closer)
Sure, the algorithm doesn't entail a neurological equivalent of what humans to classify objects, but the effect is the same, no? Is not the task of accurately classifying objects useful in and of itself? For example, how would you build said robot that can manipulate objects in a useful fashion if it cannot recognize the objects in its vicinity?
It seems that the topic of the linked article is a new unsupervised algorithm that categorizes images. The linked article says that 'the Evolution-Constructed Features algorithm is notable in that it decides for itself what features of an object are significant for identifying the object', which unsupervised algorithms do implicitly, no? It is also stated that the algorithm 'is able to learn new objects without human intervention' - so if I'm interpreting this and the article's abstract correctly, the algorithm uses a novel approach to coding some sort of more or less raw image data which it receives as input? Otherwise, it appears that what makes the approach newsworthy is its extremely high accuracy, which was 95 to 100% on some measures. That sounds very good if the tests were representative of a real-world environment.
Is this the one? It doesn't appear that the researchers have posted a manuscript, and I'm not sure that Elsevier would take kindly to it if they posted the published draft (although many researchers do so anyway). That, along with a lack of public interest in reading articles upon which pop science articles (like the one in the link) are based, probably explains the lack of a link or reference to the original article. If you have access to a library that subscribes to Pattern Recognition, you can get the article.
Anaxagoras, Socrates, Aristotle. In ancient Greece (or Athens, specifically), the charge of impiety was sometimes used for political reasons (i.e. to dispose of people that the public, or at least a few influential individuals, found to be a nuisance or an menace). Sounds like there was a political angle to the modern case as well. The more things change... Here's a link to an article about the history of this practice. That said, what was this guy trying to accomplish by mocking a dead monk? Not a cool thing to do, whatever your religious views.
A person's response to the gaming experience is not determined on a rational basis.
If a game is stimulating enough, a person will experience physiological responses that some describe as reactions to stressors - this includes a central and peripheral nervous response mediated by catecholamines (dopamine, adrenaline/norepinephrine, noradrenaline/norepinephrine) (sympathetic nervous system-adrenal-medullary arousal), and possibly pituitary-adrenal-cortical arousal, which results in a release of ACTH, and thus, cortisol. Such responses may be associated with a number of physiological effects, and influence the body's use of energy. Maladaptive psychological states, such as that of increased hostility, are sometimes associated with these changes. Here is an article which offers a pretty good introduction on the topic.
If you're prone to maladaptive responses to stressful situations, to some extent, this can be mitigated by training (hence, the biofeedback article). However, I'd be willing to guess that a lot of hardcore gamers (not all) who suffer the most severe stress effects may notice some hitches with the idea of trying to manage their stress response while gaming. Some will not be able to mitigate their stress response to a meaningful extent. Some will be able to mitigate their stress response, but it will hit eventually as they keep gaming (possibly manifesting itself pretty quickly and powerfully). Some will find that mitigating their stress response compromises or interferes with their gaming experience or their level of play, and will drop the idea altogether. Still, it's a worthwhile effort, because it has the potential to help some people.
Anyone else more concerned about the frivolous power consumption to which the "internet of things" will contribute?
Spam is a nuisance, but it can be mitigated by simple technological measures, such as spam filters (I won't get into the other security implications, which can be way more serious than spam). However, the effects arising from excessive, needless power consumption, are likely to be much more difficult to mitigate.
Gotta agree, to a large extent, with the AC above. US universities often seem to be a much more serious proposition at the grad level than at the undergrad level, although this can vary quite a bit from university to another, and from one concentration/major to another. US universities' reputations have more to do with their ability to provide a heavy duty grad (i.e. professionalizing) education and with their research output than they do with their undergrad offerings (which is often a hand-holding jog, buffeted by rampant grade inflation (lest someone not get his tenure due to somebody being upset about their grade)). Having spoken to people from Eastern Europe, I get the impression that their schools have less tolerance for sub-par performance and less grade inflation, and come exam time, you are expected to know your stuff exceptionally well.
However, as mentioned elsewhere in this thread, E. European professors tend to be underpaid (something they share with their colleagues in other countries, but it's obviously quite a bit less harsh here), which results in high levels of bribery - you can either really earn your diploma, or you can buy your diploma. Amazingly, even med schools and engineering schools seem to be susceptible to this problem.
I read the entire comment. On which "word" am I overly focused? Is it this "word":
I'm happily surprised to learn I am not the only one who thinks the whole 'gaussian' should be banished.
Am I taking that statement, which was presented in its own paragraph, out of context? Or are you arguing that you went on to contradict that statement later in your post? You understand the difference between developing a thought and contradicting it, don't you?
I've repeatedly given you the opportunity for discussion. So far, you haven't demonstrated the intellectual capacity for that. Instead, you've offered unsubstantiated assertions, accusations, and ad-hominem attacks. My post was on topic, and you're doing everything you can to avoid staying on topic, focusing instead on the fact that I reiterated a sentiment that you expressed. Instead of admitting that your post was poorly articulated or potentially self-contradictory, you offer unsubstantiated assertions that you're correct.
So far, along with slinging mud, your level of discourse has been precisely this:
I'm right about the Gaussian Distribution
So I'll give you another chance to prove that you have the intellectual capacity for something other than repeating dogma that you haven't quite integrated or understood, and for personal attacks: go ahead and respond to the points I made in my post. Stop pretending that stuff which directly addresses what you said is off-topic. Otherwise, go ahead and keep perpetuating the impression of yourself (thick, argumentative, and with a superficial understanding of things; a fraud) that you created in this thread. Otherwise, if you maintain your present course, and continue to contribute nothing of meaning or value, you can go ahead and fuck off.
The fact that you say this in response: "You categorically said that the Gaussian distribution should be banished." is bullshit and shows me that you are either a) trolling or b) not engaging with the topic enough to justify my time typing.
An exact quote from your post: "I'm happily surprised to learn I am not the only one who thinks the whole 'gaussian' should be banished".
You had no problem justifying your time typing a reply to a post that you deem unworthy of your time, so I call bullshit. So far, you've put forth the impression (and I'm sure I'm not alone in this) that you've put forth a dogmatic point of view, and that you are unable to defend that point of view. My post raised a number of valid points, none of which you've attempted to address. I understand that that it's much easier to dismiss a post as a troll without basis than to engage in rational discussion, and others who read this thread will understand that as well.
So far, I've offered a rebuttal to your anti-Gaussian point of view, and you've posted nothing of any substance. If it makes you happy, omit that phrase which you find contentious from my above post, and post your response. At this point, you're the one giving the impression of trolling (i.e. trolling those who have an understanding of the Gaussian distribution's utility in statistics and the sciences).
Gaussian distributions are not natural phenomenon. Numbers are not a natural phenomenon.
Gaussian distributions are an arithmetical description of a very common natural phenomenon. Sure, they are a man-made construct, but they are 'natural' in the sense that they describe a quality that's inherent in collections of things (although you can also argue that the notion of 'collections' or 'groups' may arguably also not be a natural phenomenon). Talking about a Gaussian distribution is about as natural as saying '1/3 of them have quality x' or talking about the circumference of a circle.
If humans want to use a Gaussian distribution to get rid of noise in some signal from a WIMP detector, fine....that's not really what we're talking about here.
You categorically said that the Gaussian distribution should be banished. Although your post focused on science, it wasn't implied in your quote that you were only talking about science.
Using a Gaussian distribution to determine how far from random your Likert scale test of whether video games make people feel more aggression...well that's ruining science.
The Likert scale thing is an example in which, in principle, the Gaussian distribution is not appropriate, and thus, should not be used. However, in practice, it typically works just fine. You can find plenty of literature for or against doing so. Here's an example in which the author argues that it's permissible.
But even if you don't want to apply parametric methods to Likert scales, what's wrong with, let's say, using Gaussian distributions to analyze participants' response times on some perceptual test? How is that, from a statistical perspective, qualitatively different from "getting rid of noise in some signal from a WIMP detector"?
What really would screw up science is the inability to utilize useful statistical tools to perform analyses - this would be the effect of "banishing the Gaussian distribution" from analyses used in scientific research.
Concerning his education credentials: he's got a U. Penn. MBA and a U. of Paris doctorate, and currently teaches at NYU Polytech. If you want to know his thoughts on normal distributions, stats, epistemology, econ, and the social sciences, his books are excellent, and are well worth a read (although much of the best material is quite derivative of Mandelbrot). NNT may be called an anti-academic, anti-econ establishment crank, but it would generally be inaccurate, in accordance with your inference, to accuse him of lying.
Data science is a field that combines machine learning and statistics to derive meaning from data. Data scientists should be reasonably well-versed in classical stats, but the data sets they deal with are often huge, ill-defined, and not amenable to analysis using classical methods. To deal with such challenges, data science recruits a healthy combination of certain areas of comp-sci (databases, machine learning, NLP, AI), statistical methods, and, quite often, improvisation.
Strange that there are so many people on here that are unfamiliar with data science.
"If people believe that they have to give up a comfortable lifestyle to reduce carbon dioxide emissions, they will look for any evidence that AGW is incorrect, no matter how flimsy it is. You can see this behavior for what it is when people cling to a mistaken idea for dear life."
The above reminded me of something from Nassim Taleb's writings. Those who have read his books may be familiar with the following Upton Sinclair quote: "It is difficult to get a man to understand something, when his salary depends upon his not understanding it." NNT applies this principle to financial 'experts' (quants, stockbrokers, advisors, etc.) who do things that are demonstrably counterproductive (applying stat methods that assume Gaussianity to non-normal distributions; disregarding the randomness inherent in stock movements) not necessarily out of ignorance, but largely because such actions serve their economic benefit. In all areas, people often disregard evidence when doing so serves what they perceive as their immediate interests.
You can say the cost is too much, but the doctors need to be paid, and their education is not cheap, so they should be paid well
Insurance companies have recently been getting creative in their attempts to keep costs down. Perhaps another thing they may want to consider is to subsidize the education of med school students who intend to start their own practices, with those individuals' practices then giving discounts/preferred treatment to the customers of the companies that subsidized them to the amount of some value exceeding the amount of financing that was provided by the insurance company (adjusted for inflation, taking into account the risk that not all med school students will start a successful practice, and so on).
Most graduate (including Ph.D.) students take a lot of classes on basics (at the start) so that they know the vocabulary and concepts necessary to read and understand the cutting edge research. Without that, you are likely too dependent on the tool.
I'd modify that to saying that without a sufficient theoretical background, even if you have access to the best software and hardware tools, you will not be able to do very much with them that will be of interest to anyone. Lots of great research (including research in the domain of computational neuroscience) is done on FOSS tools that can be downloaded in a matter of a few minutes; the prerequisite is having sufficient knowledge in a particular domain. Conversely, no software will compensate for a lack of knowledge. On a related note, I had a neuroscience professor who would say that if you're too focused on the tools you're using, you're stepping away from being a researcher, and towards being a technician.
Note: I've published cognitive and neuroscience research that utilized neural nets. I'm not specifically that knowledgeable in the specialized topics listed after point (4), but perhaps I can provide some useful general information about how to go about acquiring resources that may help the author, and perhaps others, increase their chances of success in their research efforts.
(1) What are some interesting computational neuroscience simulation problems that an individual with a workstation class PC can work on?
The first step would be to get a very solid theoretical grounding in your field. "The basic concepts behind Computational Neuroscience" is a start, but getting a good grounding in at least neuroscience and cognitive science, as well as other subdomains of psych, would help you tremendously (that's a soft way of saying that it's probably imperative to have this background). In the process of doing so, you will become familiar with research that has been done, and you will get a better idea of which specialized topics in your field appeal to you; furthermore, you will have an idea of the research that's currently being done, and which research it's reasonable for you to pursue with the specific computational and data resources that you either have at your disposal or with resources to which you can gain access.
I realize that there's a strong DIY/autodidact ethos here, but for the purposes of getting a theoretical grounding, consider enrolling in a grad program, if it's at all feasible. This approach is likely to bring you up to speed more quickly than learning things on your own (consider exploring a city which you've never visited: you can figure things out on your own, but a qualified guide or resident can show you the most important sights while wasting a minimum amount of time).
Google Scholar will be a very useful resource for you, but if you can, try to get access to a university library so that you can electronically access the journals to which they're subscribed - I understand that some very good libraries provide paid access.
(2) Is it easy for a non-academic to get the required data?
There's too many variables to provide an answer to this question. Sometimes, you get lucky and obtain data with a minimum of fuss. Sometimes, you can't obtain data at all because whatever person or organization has that data doesn't want to share with you. Sometimes, the data you may want has not even been generated yet (in which case, it's useful to have a lab, get grants, etc.).
There are some data repositories which will not give you access without some sort of qualification, such as being an NIH-funded researcher. To obtain data from these, you may want to get affiliated with a lab at a university, and ask them to obtain data for you - if you manage to convince them that you'll be a valuable asset, they may be willing to do you the favor of obtaining data from the repository, as well as sharing their own data.
Sometimes, you can get data by contacting labs directly. Needless to say, being affiliated with a university in some capacity (i.e. perhaps being a research assistant somewhere, or at least having a.edu email address) will increase your chances of obtaining data. Unfortunately, some labs will invariably not want to share their data at all.
(3) I am familiar with (but not used extensively) simulators like Neuron, Genesis etc. Other than these and Matlab, what other software should I get?
Identify the research areas that interest you. Look up papers in those areas on google Scholar (plenty of full texts are available in PDF and HTML). In the methods sections, it is conventional to state which programs the researchers used.
Speaking from personal experience, Python, Matlab, and R have served me well through providing a number of useful modules and functionalities, including neural net libraries, text processing, and statistical analysis. Don't underestimate what you can do with these basic tools and some open source librar
(**) Also, I have no meaningful training in science or statistics. If you want, you can win the argument by pointing this out in your response.
It's not my intention to get into any arguments or win anything. When I got snarky above, it was to get some people to consider whether they're qualified to disparage the work of others. Anyway, you raise a good question.
Like all sciences, psychology entails a set of beliefs/theories/ideas/models. These constructs should be informed by evidence gathered through a certain methodology. As new evidence is gathered, old models are continually revised or superceded by new ones in an iterative process that's intended to provide a better understanding of things than the understanding we had.
All sciences come up with models that are wrong or grossly oversimplified (thus, the need for scientists to reproduce and build upon others' research). However, the presence of wrongful models in a scientific field shouldn't discredit the field, and doesn't even necessarily mean that the field is headed in the wrong direction, because in principle, we will eventually replace ideas with better ones.
You've presented some findings from the field of psychiatry (which is a medical field, and not typically classed as a subfield of psychology; the medical field has its own set of complications, including pharma companies pushing researchers to pump out products) which (hopefully) show that researchers are continually revising their understanding of depression. This is comparable to, let's say, the various models of an atom that physicists came up with until they came to what we have today.
As far as psychology goes, consider the complexity of the human body, the nervous system, possible environmental inputs to the nervous system, and the potential range of human behaviors. Consider the complexity of the interactions between those systems. That complexity should give some idea of the difficulties intrinsic to the field of psychology.
Don't worry about the post getting cut off - the point you were trying to make is clear. To address the idea that "the methods are different from the physical reality of my setup" - it's simply not possible to be comprehensive in the instructions you provide when you write up your methods, and furthermore, in reality, variables are often introduced in a lab which impact experimental results, but which are not accounted for in the methods writeup because the authors are not aware of these variables (this is known to impact the likelihood that results will be reproduced). Adhering to certain methodological principles may mitigate the latter issue, but you can't be sure that the issue has been resolved.
Consider that it's believed that many older studies in psych, if conducted today, would produce non-significant p-values simply for the reason that over time, the population from which participants will be drawn has become different from the population from which participants were drawn for the original study (the changes may be cultural, but may be also be something that we cannot readily pinpoint). Methods sections cannot be comprehensive in addressing all concerns, and don't account for this sort of thing. Are the original studies failed studies? Should the conclusions of the original studies simply be revised to qualify the generalizability of these studies, and if so, at what point can we be reasonably confident that the conclusions of a study are generally applicable? So that's a commonplace scenario illustrating that the notion of being able to reproduce results is more slippery than we may have thought.
Slashdot Mirror
When is the last time you negotiated prices at the grocery store?
Every time you buy something at the grocery store, you are negotiating that item's price. If you refuse to buy an item because it's overpriced, the company that manufactures that item has the choice to either lower the price, try to market to a demographic that's willing to pay a higher price, or to attempt to keep selling an item at a price point at which the item won't sell.
By buying or not buying an item, you help determine demand for that item, thereby, setting its price. Basic stuff. Same set of principles applies to the labor market.
A few things about the reddit situation:
- reddit's staff will become less talented over time, as long as this policy persists, until they find some talent-price equilibrium that reflects reddit's salary policy.
- the CEO that implemented this policy will not suffer financially for doing so. However, it is likely that the talent of reddit's staff will decrease over time, and may result in her premature dismissal, possibly with an overly generous golden parachute.
- this salary policy has only a tenuous connection, if any, to the issue of women and men receiving an equal salary for equal labor. If reddit's CEO is not an absolute fool, the gender equality explanation is a rationalization intended to make an ill-conceived cost-cutting measure more palatable. If reddit's CEO is sincere in her explanation for this move, the board should remove her (and other companies should avoid her), because she does not understand certain fundamental labor market principles.
Here's another prediction: reddit will not save as much money as they think they will, because their HR department will be forced to dedicate more time to efforts to acquire adequate talent.
Have you met some of the MBAs who teach business courses? For a shock, try asking a few of them some fundamental stats questions that a person who has taken some grad-level stats courses (a prerequisite for many scientific/quantitative fields) should be able to answer. I can tell you about MBA profs who use statistical analysis allegedly on a regular basis without knowing the term "R^2".
Ballmer's probably a step up from quite a few people career academics in the business field.
Biz 101: How to ingratiate yourself with the right people
Biz 324: Managing technical staff for the non-technical (focus on developers)
Biz 412: Diversifying your product offerings, and what to do when you fail at it
The hard part is motivation. The problem with school is the knowledge looks "fake" to the majority of young schoolchildren, because it has no apparent connection what adults do. Small children are very quick to get excited about mimicking "real" things -- things they observe adults doing.
Agree with a notion that extends what you said there. Children have less ideas/concepts/mental schemas on which to draw, so much of what is presented to them in school appears arbitrary. One of the paths towards more effective education is to associate new knowledge with concepts that are present in a student's mind. This is pretty basic stuff that William James was saying in the 19th century, and which has a more modern instantiation in our understanding of artificial and biological neural networks. In a nutshell, when teachers present ideas, if it seems that the idea appeared out of a vacuum, or is arbitrary, the student will either be skeptical about the idea, or will have a hard time integrating it (i.e. may have to take the idea on faith and rotely memorize it, and retention will be difficult due to a lack of association with other concepts). The process of relating new ideas to ideas that have already been integrated by a person is crucial to the process of learning.
I honestly don't see how people are making it. I think the best teachers now go to private schools or colleges, and many (but not all, mind you) of the ones who remain are the ones who just aren't very good.
Correct me if I'm wrong, but I've read that many teachers who move to private schools from public institutions take a pay cut to do so, and they do it because they find the environment more comfortable and more conducive to doing their job. The students tend to be more classwork-oriented, and they have less disciplinary issues, so although there's less financial incentive to work in private schools (including no prospect of a nice pension upon retirement), the experience itself is more gratifying. Teachers in some urban areas, such as Chicago, for example, can make six figure salaries, along with some very nice perks; I'm not sure how that has worked out for them in terms of educational outcomes.
It seems that every aspect of the education experience feeds into every other aspect. Indifferent parents (as you've pointed out) produce indifferent or downright hostile kids, who, in turn, make many of the best teachers run for the hills, which leaves the parents and kids to deal with the less well-equipped teachers, and this process propagates through the generations. Most of the fix to education should will probably come with a cultural shift, rather than by throwing more technology into the classroom. People were learning trigonometry and calculus just fine before they had computers in the classroom, and there's no reason they can't do so now. Use computers when necessary (i.e. working with complex regression equations in a Stats course), and leave them out of the classroom when they're simply expedient or a hindrance. Things such as kindles, which present a great variety of books to students, can be extremely useful. Things like iPads are attention hogs, and wire young brains in ways that may not help them in their future endeavors. Use technology as appropriate, but don't pretend that it will be a panacea.
The unsupervised algorithm discussed in the article seems to code some sort of visual input and, I'd infer, to perform clustering, which permits it to assign labels (i.e. let's say, 'tree', 'human', etc.) to objects it has encountered. It can use this schema which it has constructed to assign objects it hasn't seen before to a cluster - that is, it labels novel inputs in accordance with its schema. Thus, the algorithm 'recognizes' classes of objects. I'd imagine if you granularize the detail level of detail to which the algorithm pays attention, it can recognize (categories of one object)/particular instances of a general object category/a particular object.
Sitting on a street, a "bicycle" is an object because it is most like to be operated on as a unit. But to a bicycle mechanic, a bicycle is a collection of objects, such as a frame, a seat.. and so on because they need to decompose the "bicycle" construct to do their job. To somebody on an assembly line putting together bicycle seats, a seat is (at least initially) several different objects.
That goes way beyond the task of 'classifying objects', which is what the algorithm is intended to do. It's like expecting a mechanic to drive your vehicle for you once he has repaired it.
So, truly unsupervised algorithms cannot do useful recognition - that is, classify objects the same way people do. (A robot that could experiment with its environment and learn to use "objects" could come closer)
Sure, the algorithm doesn't entail a neurological equivalent of what humans to classify objects, but the effect is the same, no? Is not the task of accurately classifying objects useful in and of itself? For example, how would you build said robot that can manipulate objects in a useful fashion if it cannot recognize the objects in its vicinity?
First of all, is this the right paper?
It seems that the topic of the linked article is a new unsupervised algorithm that categorizes images. The linked article says that 'the Evolution-Constructed Features algorithm is notable in that it decides for itself what features of an object are significant for identifying the object', which unsupervised algorithms do implicitly, no? It is also stated that the algorithm 'is able to learn new objects without human intervention' - so if I'm interpreting this and the article's abstract correctly, the algorithm uses a novel approach to coding some sort of more or less raw image data which it receives as input? Otherwise, it appears that what makes the approach newsworthy is its extremely high accuracy, which was 95 to 100% on some measures. That sounds very good if the tests were representative of a real-world environment.
Is this the one? It doesn't appear that the researchers have posted a manuscript, and I'm not sure that Elsevier would take kindly to it if they posted the published draft (although many researchers do so anyway). That, along with a lack of public interest in reading articles upon which pop science articles (like the one in the link) are based, probably explains the lack of a link or reference to the original article. If you have access to a library that subscribes to Pattern Recognition, you can get the article.
Anaxagoras, Socrates, Aristotle. In ancient Greece (or Athens, specifically), the charge of impiety was sometimes used for political reasons (i.e. to dispose of people that the public, or at least a few influential individuals, found to be a nuisance or an menace). Sounds like there was a political angle to the modern case as well. The more things change... Here's a link to an article about the history of this practice. That said, what was this guy trying to accomplish by mocking a dead monk? Not a cool thing to do, whatever your religious views.
A person's response to the gaming experience is not determined on a rational basis.
If a game is stimulating enough, a person will experience physiological responses that some describe as reactions to stressors - this includes a central and peripheral nervous response mediated by catecholamines (dopamine, adrenaline/norepinephrine, noradrenaline/norepinephrine) (sympathetic nervous system-adrenal-medullary arousal), and possibly pituitary-adrenal-cortical arousal, which results in a release of ACTH, and thus, cortisol. Such responses may be associated with a number of physiological effects, and influence the body's use of energy. Maladaptive psychological states, such as that of increased hostility, are sometimes associated with these changes. Here is an article which offers a pretty good introduction on the topic.
If you're prone to maladaptive responses to stressful situations, to some extent, this can be mitigated by training (hence, the biofeedback article). However, I'd be willing to guess that a lot of hardcore gamers (not all) who suffer the most severe stress effects may notice some hitches with the idea of trying to manage their stress response while gaming. Some will not be able to mitigate their stress response to a meaningful extent. Some will be able to mitigate their stress response, but it will hit eventually as they keep gaming (possibly manifesting itself pretty quickly and powerfully). Some will find that mitigating their stress response compromises or interferes with their gaming experience or their level of play, and will drop the idea altogether. Still, it's a worthwhile effort, because it has the potential to help some people.
Cold, and not cool. A bit ironic...
Anyone else more concerned about the frivolous power consumption to which the "internet of things" will contribute?
Spam is a nuisance, but it can be mitigated by simple technological measures, such as spam filters (I won't get into the other security implications, which can be way more serious than spam). However, the effects arising from excessive, needless power consumption, are likely to be much more difficult to mitigate.
Gotta agree, to a large extent, with the AC above. US universities often seem to be a much more serious proposition at the grad level than at the undergrad level, although this can vary quite a bit from university to another, and from one concentration/major to another. US universities' reputations have more to do with their ability to provide a heavy duty grad (i.e. professionalizing) education and with their research output than they do with their undergrad offerings (which is often a hand-holding jog, buffeted by rampant grade inflation (lest someone not get his tenure due to somebody being upset about their grade)). Having spoken to people from Eastern Europe, I get the impression that their schools have less tolerance for sub-par performance and less grade inflation, and come exam time, you are expected to know your stuff exceptionally well.
However, as mentioned elsewhere in this thread, E. European professors tend to be underpaid (something they share with their colleagues in other countries, but it's obviously quite a bit less harsh here), which results in high levels of bribery - you can either really earn your diploma, or you can buy your diploma. Amazingly, even med schools and engineering schools seem to be susceptible to this problem.
I read the entire comment. On which "word" am I overly focused? Is it this "word":
I'm happily surprised to learn I am not the only one who thinks the whole 'gaussian' should be banished.
Am I taking that statement, which was presented in its own paragraph, out of context? Or are you arguing that you went on to contradict that statement later in your post? You understand the difference between developing a thought and contradicting it, don't you?
I've repeatedly given you the opportunity for discussion. So far, you haven't demonstrated the intellectual capacity for that. Instead, you've offered unsubstantiated assertions, accusations, and ad-hominem attacks. My post was on topic, and you're doing everything you can to avoid staying on topic, focusing instead on the fact that I reiterated a sentiment that you expressed. Instead of admitting that your post was poorly articulated or potentially self-contradictory, you offer unsubstantiated assertions that you're correct.
So far, along with slinging mud, your level of discourse has been precisely this:
I'm right about the Gaussian Distribution
So I'll give you another chance to prove that you have the intellectual capacity for something other than repeating dogma that you haven't quite integrated or understood, and for personal attacks: go ahead and respond to the points I made in my post. Stop pretending that stuff which directly addresses what you said is off-topic. Otherwise, go ahead and keep perpetuating the impression of yourself (thick, argumentative, and with a superficial understanding of things; a fraud) that you created in this thread. Otherwise, if you maintain your present course, and continue to contribute nothing of meaning or value, you can go ahead and fuck off.
The fact that you say this in response: "You categorically said that the Gaussian distribution should be banished." is bullshit and shows me that you are either a) trolling or b) not engaging with the topic enough to justify my time typing.
An exact quote from your post: "I'm happily surprised to learn I am not the only one who thinks the whole 'gaussian' should be banished".
You had no problem justifying your time typing a reply to a post that you deem unworthy of your time, so I call bullshit. So far, you've put forth the impression (and I'm sure I'm not alone in this) that you've put forth a dogmatic point of view, and that you are unable to defend that point of view. My post raised a number of valid points, none of which you've attempted to address. I understand that that it's much easier to dismiss a post as a troll without basis than to engage in rational discussion, and others who read this thread will understand that as well.
So far, I've offered a rebuttal to your anti-Gaussian point of view, and you've posted nothing of any substance. If it makes you happy, omit that phrase which you find contentious from my above post, and post your response. At this point, you're the one giving the impression of trolling (i.e. trolling those who have an understanding of the Gaussian distribution's utility in statistics and the sciences).
Gaussian distributions are not natural phenomenon. Numbers are not a natural phenomenon.
Gaussian distributions are an arithmetical description of a very common natural phenomenon. Sure, they are a man-made construct, but they are 'natural' in the sense that they describe a quality that's inherent in collections of things (although you can also argue that the notion of 'collections' or 'groups' may arguably also not be a natural phenomenon). Talking about a Gaussian distribution is about as natural as saying '1/3 of them have quality x' or talking about the circumference of a circle.
If humans want to use a Gaussian distribution to get rid of noise in some signal from a WIMP detector, fine....that's not really what we're talking about here.
You categorically said that the Gaussian distribution should be banished. Although your post focused on science, it wasn't implied in your quote that you were only talking about science.
Using a Gaussian distribution to determine how far from random your Likert scale test of whether video games make people feel more aggression...well that's ruining science.
The Likert scale thing is an example in which, in principle, the Gaussian distribution is not appropriate, and thus, should not be used. However, in practice, it typically works just fine. You can find plenty of literature for or against doing so. Here's an example in which the author argues that it's permissible.
But even if you don't want to apply parametric methods to Likert scales, what's wrong with, let's say, using Gaussian distributions to analyze participants' response times on some perceptual test? How is that, from a statistical perspective, qualitatively different from "getting rid of noise in some signal from a WIMP detector"?
What really would screw up science is the inability to utilize useful statistical tools to perform analyses - this would be the effect of "banishing the Gaussian distribution" from analyses used in scientific research.
"I'm happily surprised to learn I am not the only one who thinks the whole 'gaussian' should be banished."
How exactly do you banish a probability distribution? Should the Central Limit Theorem be discarded along with it?
Concerning his education credentials: he's got a U. Penn. MBA and a U. of Paris doctorate, and currently teaches at NYU Polytech. If you want to know his thoughts on normal distributions, stats, epistemology, econ, and the social sciences, his books are excellent, and are well worth a read (although much of the best material is quite derivative of Mandelbrot). NNT may be called an anti-academic, anti-econ establishment crank, but it would generally be inaccurate, in accordance with your inference, to accuse him of lying.
Data science is a field that combines machine learning and statistics to derive meaning from data. Data scientists should be reasonably well-versed in classical stats, but the data sets they deal with are often huge, ill-defined, and not amenable to analysis using classical methods. To deal with such challenges, data science recruits a healthy combination of certain areas of comp-sci (databases, machine learning, NLP, AI), statistical methods, and, quite often, improvisation.
Strange that there are so many people on here that are unfamiliar with data science.
"If people believe that they have to give up a comfortable lifestyle to reduce carbon dioxide emissions, they will look for any evidence that AGW is incorrect, no matter how flimsy it is. You can see this behavior for what it is when people cling to a mistaken idea for dear life."
The above reminded me of something from Nassim Taleb's writings. Those who have read his books may be familiar with the following Upton Sinclair quote: "It is difficult to get a man to understand something, when his salary depends upon his not understanding it." NNT applies this principle to financial 'experts' (quants, stockbrokers, advisors, etc.) who do things that are demonstrably counterproductive (applying stat methods that assume Gaussianity to non-normal distributions; disregarding the randomness inherent in stock movements) not necessarily out of ignorance, but largely because such actions serve their economic benefit. In all areas, people often disregard evidence when doing so serves what they perceive as their immediate interests.
You can say the cost is too much, but the doctors need to be paid, and their education is not cheap, so they should be paid well
Insurance companies have recently been getting creative in their attempts to keep costs down. Perhaps another thing they may want to consider is to subsidize the education of med school students who intend to start their own practices, with those individuals' practices then giving discounts/preferred treatment to the customers of the companies that subsidized them to the amount of some value exceeding the amount of financing that was provided by the insurance company (adjusted for inflation, taking into account the risk that not all med school students will start a successful practice, and so on).
Most graduate (including Ph.D.) students take a lot of classes on basics (at the start) so that they know the vocabulary and concepts necessary to read and understand the cutting edge research. Without that, you are likely too dependent on the tool.
I'd modify that to saying that without a sufficient theoretical background, even if you have access to the best software and hardware tools, you will not be able to do very much with them that will be of interest to anyone. Lots of great research (including research in the domain of computational neuroscience) is done on FOSS tools that can be downloaded in a matter of a few minutes; the prerequisite is having sufficient knowledge in a particular domain. Conversely, no software will compensate for a lack of knowledge. On a related note, I had a neuroscience professor who would say that if you're too focused on the tools you're using, you're stepping away from being a researcher, and towards being a technician.
Note: I've published cognitive and neuroscience research that utilized neural nets. I'm not specifically that knowledgeable in the specialized topics listed after point (4), but perhaps I can provide some useful general information about how to go about acquiring resources that may help the author, and perhaps others, increase their chances of success in their research efforts.
(1) What are some interesting computational neuroscience simulation problems that an individual with a workstation class PC can work on?
The first step would be to get a very solid theoretical grounding in your field. "The basic concepts behind Computational Neuroscience" is a start, but getting a good grounding in at least neuroscience and cognitive science, as well as other subdomains of psych, would help you tremendously (that's a soft way of saying that it's probably imperative to have this background). In the process of doing so, you will become familiar with research that has been done, and you will get a better idea of which specialized topics in your field appeal to you; furthermore, you will have an idea of the research that's currently being done, and which research it's reasonable for you to pursue with the specific computational and data resources that you either have at your disposal or with resources to which you can gain access.
I realize that there's a strong DIY/autodidact ethos here, but for the purposes of getting a theoretical grounding, consider enrolling in a grad program, if it's at all feasible. This approach is likely to bring you up to speed more quickly than learning things on your own (consider exploring a city which you've never visited: you can figure things out on your own, but a qualified guide or resident can show you the most important sights while wasting a minimum amount of time).
Google Scholar will be a very useful resource for you, but if you can, try to get access to a university library so that you can electronically access the journals to which they're subscribed - I understand that some very good libraries provide paid access.
(2) Is it easy for a non-academic to get the required data?
There's too many variables to provide an answer to this question. Sometimes, you get lucky and obtain data with a minimum of fuss. Sometimes, you can't obtain data at all because whatever person or organization has that data doesn't want to share with you. Sometimes, the data you may want has not even been generated yet (in which case, it's useful to have a lab, get grants, etc.).
There are some data repositories which will not give you access without some sort of qualification, such as being an NIH-funded researcher. To obtain data from these, you may want to get affiliated with a lab at a university, and ask them to obtain data for you - if you manage to convince them that you'll be a valuable asset, they may be willing to do you the favor of obtaining data from the repository, as well as sharing their own data.
Sometimes, you can get data by contacting labs directly. Needless to say, being affiliated with a university in some capacity (i.e. perhaps being a research assistant somewhere, or at least having a .edu email address) will increase your chances of obtaining data. Unfortunately, some labs will invariably not want to share their data at all.
(3) I am familiar with (but not used extensively) simulators like Neuron, Genesis etc. Other than these and Matlab, what other software should I get?
Identify the research areas that interest you. Look up papers in those areas on google Scholar (plenty of full texts are available in PDF and HTML). In the methods sections, it is conventional to state which programs the researchers used.
Speaking from personal experience, Python, Matlab, and R have served me well through providing a number of useful modules and functionalities, including neural net libraries, text processing, and statistical analysis. Don't underestimate what you can do with these basic tools and some open source librar
(**) Also, I have no meaningful training in science or statistics. If you want, you can win the argument by pointing this out in your response.
It's not my intention to get into any arguments or win anything. When I got snarky above, it was to get some people to consider whether they're qualified to disparage the work of others. Anyway, you raise a good question.
Like all sciences, psychology entails a set of beliefs/theories/ideas/models. These constructs should be informed by evidence gathered through a certain methodology. As new evidence is gathered, old models are continually revised or superceded by new ones in an iterative process that's intended to provide a better understanding of things than the understanding we had.
All sciences come up with models that are wrong or grossly oversimplified (thus, the need for scientists to reproduce and build upon others' research). However, the presence of wrongful models in a scientific field shouldn't discredit the field, and doesn't even necessarily mean that the field is headed in the wrong direction, because in principle, we will eventually replace ideas with better ones.
You've presented some findings from the field of psychiatry (which is a medical field, and not typically classed as a subfield of psychology; the medical field has its own set of complications, including pharma companies pushing researchers to pump out products) which (hopefully) show that researchers are continually revising their understanding of depression. This is comparable to, let's say, the various models of an atom that physicists came up with until they came to what we have today.
As far as psychology goes, consider the complexity of the human body, the nervous system, possible environmental inputs to the nervous system, and the potential range of human behaviors. Consider the complexity of the interactions between those systems. That complexity should give some idea of the difficulties intrinsic to the field of psychology.
Don't worry about the post getting cut off - the point you were trying to make is clear. To address the idea that "the methods are different from the physical reality of my setup" - it's simply not possible to be comprehensive in the instructions you provide when you write up your methods, and furthermore, in reality, variables are often introduced in a lab which impact experimental results, but which are not accounted for in the methods writeup because the authors are not aware of these variables (this is known to impact the likelihood that results will be reproduced). Adhering to certain methodological principles may mitigate the latter issue, but you can't be sure that the issue has been resolved.
Consider that it's believed that many older studies in psych, if conducted today, would produce non-significant p-values simply for the reason that over time, the population from which participants will be drawn has become different from the population from which participants were drawn for the original study (the changes may be cultural, but may be also be something that we cannot readily pinpoint). Methods sections cannot be comprehensive in addressing all concerns, and don't account for this sort of thing. Are the original studies failed studies? Should the conclusions of the original studies simply be revised to qualify the generalizability of these studies, and if so, at what point can we be reasonably confident that the conclusions of a study are generally applicable? So that's a commonplace scenario illustrating that the notion of being able to reproduce results is more slippery than we may have thought.