but Prof. Andrew Ng says in his video for the machine learning class "If you successfully complete this class you also get from me a signed statement of accomplishment stating how you did on the class that you can put on your resume." that's got to mean something to some people given his reputation in the field. Especially those who are trying to scoop up as many machine learners as possible in this whole Big Data rush.
You have no idea what you're talking about. The abstract clearly illustrates their type I probability as at most.05 which is pretty standard.
Using other figures from the abstract I approximate their type II probability using the following
approximate critical t value (from the t-table in Wackerly's Mathematical Statistics) 1.96 for 46 df
t confidence interval
4.2 = 2.435 + 1.96*se
implies se =.9005
since se = std.dev./sqrt(n) (recall n=47)
std. dev. = 6.1736 (approximately)
The effect size (d) is approximately.3934 under equal variance
This leads to the following approximate power
For a One-Tailed (Directional) Hypothesis
Observed Power: 0.778
For a Two-Tailed (Non-Directional) Hypothesis
Observed Power: 0.671
They had a directional hypothesis (mu phone on greater than mu phone off) and since I'm only using approximate values, numbers from their abstract, and I assumed by their use of ANOVA that equal variance was satisfied (so I used the same value for both) that that.778 is pretty close to the standard.8. So I'm willing to bet their experimental design was such that it their power is actually at least.8 and that the difference is from my rounding and approximation and not actually using their recorded standard deviations (since I didn't see them in the abstract and I don't have access to the JAMA article itself).
If any methodological challenge can be had from just reading an abstract (which seems unlikely since all of their methods are not expressed in it) its that they don't explain which version of ANOVA they used (it would seem repeated measures is most appropriate and a quick review of other studies suggests they likely used this, but since they don't say it explicitly its possible they used another ANOVA). One might also question why a regression model with mixed effects wasn't used given the repeated measures and potential for significant mixed effects in longitudinal study, but given the relationships between ANOVA and regression it isn't really a fatal flaw.
Now posting that ridiculous (see *) Science News article could only have one of a few purposes: (1) You're a Bayesian and you're rejecting a frequentist approach to this study. (2) You're against the use of all statistics based on their uncertainty versus deterministic/certain models from mathematics. (3) You think compounding errors effects this study. If I'm missing you're real point please feel free to elaborate it.
If (1) I'm not going to dive into the pro/con's of Bayesian vs Frequentist, but I will say Bayesian models are built off of Frequentist its not like they were independently developed in a vacuum. As such while Bayesian methods act generally under different assumptions they do inherit and are confined by some of the same restrictions as frequentist models. A real problem is that Bayesian models work well when you can effectively incorporate prior knowledge (as in domain specific knowledge) which a Statistician isn't likely to have, and how realistic do you really think it is to teach scientists Bayesian methods when they all require some understanding of frequentist models that they already have shown they don't fully understand?
If (2) well if you have a neat proof for PvsNP that P=NP tucked away you might want to get on with submitting it and claiming your million dollar prize and possible fields medal if not other accolades or if you have many previously unpublished exact methods please publish them we're at a point where computing power could actually do the necessary calculation and that too could possibly net you a fields medal and more. Since most methods used are based on maximum likelihood, most powerful test, or some other "at least this good" type method you don't need to fear the uncertainty and without a way to map probabilistic methods to deterministic ones meaningfully an
Would a 2.4 micromole per minute change in glucose metabolism in the orbitofrontal cortex and temporal pole region be of any practical significance? What is the expected value and what is considered "normal" (perhaps not in the statistical sense) variation for glucose metabolism you'd see in a PET scan in this part of the brain in general?
I get that the study shows that it is statistically significant (they use a two sample t-test and some version of ANOVA for multiple comparisons, hopefully they used repeated measures ANOVA since that's more appropriate (but maybe regression with mixed effects would be even more appropriate still) since at least some subjects are in both groups by their randomized crossover design.
I'm just curious if this is a case of a result that is statistically significant, but not really of any practical significance or if 2.4 micromoles per minute change from expectation would be something that alarmed a Neurologist after looking at a PET scan for this region.
I'm pretty sure a large and dense neural network even using something as well known and "simple" as backpropagation can generate a net that's not comprehensible to the human's that programmed it.
This will probably burn karma points, but I have a strong feeling this needs to be said.
On some level your statement that "If you can't do (or follow) the math, then you don't know the physics" is true, but only in the case of a reduction to the absurd. I'm talking about someone who literally has no grasp of any basic number skill, geometry concept, or ability to observe naturally occurring macroscopic phenomenon. In that case though I think there needs to be a reasonable explanation as to why such a person would even be interested in or exposed to the topic to begin with?
"Physics IS mathematics" - more like physics is EXPLAINED with mathematics. I fixed that for you.
You don't need to know how to compute an integral or moreover know the details of the underpinning logic statement represented by an integral expression or how to prove one mathematically (not evaluate one) to appreciate the generalization that integrals measure areas, volumes, and higher dimensional analogies.
You don't need to know the chain rule, the general power rule, the product or quotient rule let alone the terribly complicated logical statements underpinning the notation to appreciate the generalization that derivatives are rates of change.
You don't need to know how to work with vector valued functions to understand that they can represent the objects and their attributes both macro and micro in a wide variety of circumstances.
You don't need to know how to express mathematics as a provable logical statements in order to appreciate that it works. Being able to evaluate an expression is not the same thing as proving the fundamental underpinnings that expression is talking about.
Physicists aren't generally sitting around proving new theorems (unless you're Witten, Werner, maybe a couple of others, or some of their grad students) and publishing them in math journals. Usually physicists are implementing an instance of some already established area of mathematics that seems to suit their theoretical needs so they can predict something and go look for it in a lab. Or they are just asking mathematicians for assistance like Weber and Gauss which is neither the first or last of such an academic coupling.
The point is most physics models are implementations of existing maths under the assumption that the math is logically sound and meaningfully applies to what they want to know. Such as using an integral to measure energy stored in capacitance. That's a task much more akin to computer programming than it is to the activity of mathematicians.
If someone can't adequately break down what they are doing to help someone who knows very little about the subject matter "get it" or at least appreciate its marvel on some level without following through the mechanics of some mathematical algorithm than I'd be worried that that person doesn't really know what they're doing and is resorting to "blue box" strategies.
Thought experiments (or summaries on scientific research and problems -- how do you expect to attract young scientists who aren't reading academic journals because they don't have access or are too young?) have been crucial to the development of the sciences so the outright dismissal of their merits because there is some sentiment that there should be some necessary technical threshold to be met before one can contribute shows a complete lack of understanding or appreciation of how much such discourse has progressed our understanding of mathematical and scientific principles. Even if it means addressing silly questions seriously sometimes.
Well they are not one in the same sure, but The Swedish Pirate Party also hosts The Pirate Bay itself so you can't completely separate them from each other either.
The Pirate Party Becomes The Pirate Bay’s New Host
Obviously both sites and the Swedish Pirate Party are betting (pretty hard) on the election next month which a successful outcome would as previously posted put TPB and perhaps now wikileaks inside the Swedish Parliament.
They probably are relying on or hoping to attain the same standing RIAA has gotten through SoundExchange to collect broadcast royalties even from non-members.
So there is clearly precedent that suggests ownership and membership are not sufficient concerns to these types of organizations. Unless it is their material or members that is!
So in this case they are either seeking statute authority to collect song composing royalties for members AND non-members, or they intend to behave that way anyway and defend it on the premise that the copyright office already delineated similar powers to SoundExchange and that since ASCAP is a similar group to SoundExchange they are entitled to a similar wide scope of authority (performance royalties -> SoundExchange vs composing royalties -> ASCAP)
I'd really like to see this blow up in their face and get both groups rights to even try this sort of thing revoked, but there are too many MAFIAA members in DoJ (and probably other parts of gov't) now and they have the administration's support (much to my dismay as I do generally otherwise support the administration). So this could get ugly and have bad consequences quickly.
I really hope the copyleft groups start gathering funds and resources in a way to respond to this head on. I'd support it.
Synthetic biology has two paradigms the first is the top down approach which deals with gene knock outs to look for minimal sets necessary for life that can then be tailored to suit specific needs/tasks. The other approach is the bottom up which have inspiration from the Miller-Urey experiments. They are trying to spontaneously generate a new biological system from scratch. Some researchers in this camp trying to create synthetic cellular components in hopes of putting all the synthetic parts together to create a functioning cell such as synthetic Golgi bodies
There has been some promising results from both approaches. It is a pretty exciting time in Biology.
I've seen many many posts talking about "you need logic but not really mathematics"
I'm not really sure what the confusion is about why mathematics would be separate from logic in any respect.
Symbolic Logic is Mathematical Logic. Principia Mathematica clearly makes the case that math and logic are the same thing. As did many works around the same time and since. So how can you need logic for computer science or programming yet somehow not mathematics it just doesn't make any sense to me.
As Dvorkin put it, it isn't as simple as always use classical methods or always use Bayesian.
I also think sometimes whenever possible you need to use both to help give you more evidence. Particularly if the results are somehow unclear. If they both give similar or approximately the same results that might be pretty good evidence that the final result is reasonable.
If you get wildly different conclusions then you need to consider why you got different answers and maybe it is because one method better models the situation than the other (and there are some tools to help figure that out in some cases) or perhaps the irregularity is identifying a problem with the data collected that would have gone undetected if you only used one or the other (and there are tools for when this happens too in some cases).
Very careful model building, very careful experimental design, very careful analysis, and very careful model adjustment isn't a guarantee that the results will be correct with any methods, but it is the best that can be done and is critical to always follow.
Definitely learn to use R since its free you don't have to worry about paying licensing fees. It is also widely used (no matter what you here from SAS, Minitab, SPSS, etc).
Devore talks about Bayes Rule as does Wackerly and Wackerly's last chapter talks about some Bayesian techniques, but these are merely primers for what is typical in a Bayesian course. So I recommend these two books as analogous with Devore's: Bolstad Introduction to Bayesian Statistics and to Wackerly's: Hoff A First Course in Bayesian Statistical Methods
Some things you need from mathematics are the ability to integrate, work with matrices and matrix operations, and algebraic manipulation. Familiarity with transformations and operators especially linear ones is useful since many procedures in statistics are linear operators. The highest levels of statistics will get even more math intense using mathematical results from areas like ODE/PDE, Galios Theory, or general Measure Theory.
The wikipedia's statistics articles are pretty good overall, but as Dvorkin noted some are more technical than what would be friendly to those that are new to statistics. When you feel that's the case try using the sources linked as citations in the article or google confusing parts and it is generally possible to find an explanation for almost any background level.
However if you can get through these texts you're background would be pretty strong.
It is not a shortcoming of statistics that other people, like various scientists who aren't statisticians, don't know how to use or properly interpret statistics. It is a shortcoming of their knowledge.
It is not a shortcoming of the Copenhagen interpretation of quantum mechanics or the Chicago school of economics if I don't understand or know how to correctly interpret their results. It is my shortcoming and fault for not knowing enough to connect the dots.
I do statistical research some of that is through interacting with researchers in the biosciences. Often when I go to talk to a researcher and ask them if they could use some statistical or mathematical or computational assistance with their research it has almost always been a fruitful starting point to long conversations and getting into the research. Now sometimes it was simply a matter of looking at their F-test results or ANOVA scores and telling them what it meant (like with a regression model relating proportions of certain characteristics between taxa), more useful interactions for me often mean working on new algorithms or estimators or working with fitting a model from their empirical data because there isn't a reliable standard model to work off of (like intergenic distance between genes in an operon) that kind of challenge makes less engaging work worth the hassle. Maybe I'm odd because I've worked hard to have a good background in both statistics and biology, but I shouldn't be.
Although here is an observation that perhaps supports some of the intent of the article from my own experience. I was speaking with a biology graduate student and it came up that they had a biostatistics course in the department. Of course as a statistician my mind goes towards survival function, failure rate, life tables, censored data, bioassy, epidemiology, microarrays, clincal trials, topics along those lines. It turned out their course focused z tests, t tests, f tests, confidence intervals, point predictions, least squares regression, multiple regression, ANOVA, and things along these lines just with simulated problems in a lab setting. That is not necessarily a bad thing, but much of the core math was under played or missing like model assumptions and alternate formulations or things like dummy variables. The worst part was that even though they were doing well with the class they had no confidence in actually using the statistics and didn't understand how to interpret the meaning of something like a confidence interval, they knew how to calculate one, but it wasn't clear what it actually meant to them.
The corollary to the notion in the summary I'd rant and claim is that scientists overall have less than desirable skills in mathematics, statistics, and computation than those who studied those disciplines principally and that's hurting science. However many in those three disciplines really know little beyond basic results in any of the sciences which hurts the applicability of these mathematical fields to the sciences and likely hurt our ability to develop certain types of discipline specific results that can be generalized from work in application problems.
In either case whether you're a typical scientist or a typical math/stat/comp person in order to become proficient enough in the other areas it requires going an awfully long out of the way compared to any counterpart who simply does not care and goes straight through as many before have. While in some areas of research on either side it is no problem to do as has been done and not further knowledge into those other areas. Increasingly results that have the highest levels of impact are coming more and more from truly interdisciplinary research. In order to further encourage that for those who are interested in such fields (aside from making more clear what areas in any of the fields fringe to such interdisciplinary work) we need more incentive to study more than one field and/or better ways of enabling fruitful cooperation between the camps.
EMI and other Cope programs have MIDI output capability. He'd pipe the midi files to a program that allowed for printing as sheet music (I don't use apple's so I am not familiar with what program specifically he was sending the midi to for sheet music printing). The EMI described in Cope's books required manually keying in entries for the databases, but it would seem Zenph's only difference is that it is working with the audio capture directly. I'm pretty positive EMI or the other Cope programs could readily be adapted to do that. Also if EMI were only given a few pieces or a single piece in the analysis db and the parameters are set in a certain way EMI would do exactly what Zenph does.
I agree though that Zenph seems to want to head toward what EMI is already capable of. This gets compounded when you consider the style signatures or genetic analogies used by Cope or even a system like Pandora (which has been my impressions about what Recombinant might be all about) then it makes Zenph all the more trivial, but it points towards the increasing trend in this section of industry trying to do this sort of algorithmic composition, re-arrangement, remastering, etc.
I don't think that's a correct interpretation of copyright law.
"In the case of works made for hire, the employer and not the employee is considered to be the author." from LOC copyright circular
So if work for hire allows for corporations to create and author copyright materials then why wouldn't a corporation be able to author the copyright of the output of this sort of program?
David Cope's Experiments in Musical Intelligence and related works (SARA, other works, and his own company called Recombinant inc ) have been doing this for many years.
SIAM (Society for Industrial and Applied Mathematics) has several special interest groups related to computing/programming problems. The other major math and stat groups have excellent articles on computing problems from time to time as well like the AMS, MAA, or AMSTAT, but SIAM probably provides the most of these groups and a lot of coverage that compliments IEEE and ACM. Also depending on if you're working in a specific industry or if you're furthering your studies in a graduate program there may be other professional societies that deal with informatics or computational issues related to that focus.
Slashdot Mirror
That's 1507 systems to you and he's 11! ;)
but Prof. Andrew Ng says in his video for the machine learning class "If you successfully complete this class you also get from me a signed statement of accomplishment stating how you did on the class that you can put on your resume." that's got to mean something to some people given his reputation in the field. Especially those who are trying to scoop up as many machine learners as possible in this whole Big Data rush.
Except wasn't Tesla the victim of such tactics courtesy of Edison?
Did MS have similar restrictions placed on it when it bought Farecast?
You have no idea what you're talking about. The abstract clearly illustrates their type I probability as at most .05 which is pretty standard.
.9005
.3934 under equal variance
.778 is pretty close to the standard .8. So I'm willing to bet their experimental design was such that it their power is actually at least .8 and that the difference is from my rounding and approximation and not actually using their recorded standard deviations (since I didn't see them in the abstract and I don't have access to the JAMA article itself).
Using other figures from the abstract I approximate their type II probability using the following
approximate critical t value (from the t-table in Wackerly's Mathematical Statistics) 1.96 for 46 df
t confidence interval
4.2 = 2.435 + 1.96*se
implies se =
since se = std.dev./sqrt(n) (recall n=47)
std. dev. = 6.1736 (approximately)
The effect size (d) is approximately
This leads to the following approximate power
For a One-Tailed (Directional) Hypothesis
Observed Power: 0.778
For a Two-Tailed (Non-Directional) Hypothesis
Observed Power: 0.671
They had a directional hypothesis (mu phone on greater than mu phone off) and since I'm only using approximate values, numbers from their abstract, and I assumed by their use of ANOVA that equal variance was satisfied (so I used the same value for both) that that
If any methodological challenge can be had from just reading an abstract (which seems unlikely since all of their methods are not expressed in it) its that they don't explain which version of ANOVA they used (it would seem repeated measures is most appropriate and a quick review of other studies suggests they likely used this, but since they don't say it explicitly its possible they used another ANOVA). One might also question why a regression model with mixed effects wasn't used given the repeated measures and potential for significant mixed effects in longitudinal study, but given the relationships between ANOVA and regression it isn't really a fatal flaw.
Now posting that ridiculous (see *) Science News article could only have one of a few purposes: (1) You're a Bayesian and you're rejecting a frequentist approach to this study. (2) You're against the use of all statistics based on their uncertainty versus deterministic/certain models from mathematics. (3) You think compounding errors effects this study. If I'm missing you're real point please feel free to elaborate it.
If (1) I'm not going to dive into the pro/con's of Bayesian vs Frequentist, but I will say Bayesian models are built off of Frequentist its not like they were independently developed in a vacuum. As such while Bayesian methods act generally under different assumptions they do inherit and are confined by some of the same restrictions as frequentist models. A real problem is that Bayesian models work well when you can effectively incorporate prior knowledge (as in domain specific knowledge) which a Statistician isn't likely to have, and how realistic do you really think it is to teach scientists Bayesian methods when they all require some understanding of frequentist models that they already have shown they don't fully understand?
If (2) well if you have a neat proof for PvsNP that P=NP tucked away you might want to get on with submitting it and claiming your million dollar prize and possible fields medal if not other accolades or if you have many previously unpublished exact methods please publish them we're at a point where computing power could actually do the necessary calculation and that too could possibly net you a fields medal and more. Since most methods used are based on maximum likelihood, most powerful test, or some other "at least this good" type method you don't need to fear the uncertainty and without a way to map probabilistic methods to deterministic ones meaningfully an
Do any Neurologists read /. ?
Would a 2.4 micromole per minute change in glucose metabolism in the orbitofrontal cortex and temporal pole region be of any practical significance? What is the expected value and what is considered "normal" (perhaps not in the statistical sense) variation for glucose metabolism you'd see in a PET scan in this part of the brain in general?
I get that the study shows that it is statistically significant (they use a two sample t-test and some version of ANOVA for multiple comparisons, hopefully they used repeated measures ANOVA since that's more appropriate (but maybe regression with mixed effects would be even more appropriate still) since at least some subjects are in both groups by their randomized crossover design.
I'm just curious if this is a case of a result that is statistically significant, but not really of any practical significance or if 2.4 micromoles per minute change from expectation would be something that alarmed a Neurologist after looking at a PET scan for this region.
I'm pretty sure a large and dense neural network even using something as well known and "simple" as backpropagation can generate a net that's not comprehensible to the human's that programmed it.
This will probably burn karma points, but I have a strong feeling this needs to be said.
On some level your statement that "If you can't do (or follow) the math, then you don't know the physics" is true, but only in the case of a reduction to the absurd. I'm talking about someone who literally has no grasp of any basic number skill, geometry concept, or ability to observe naturally occurring macroscopic phenomenon. In that case though I think there needs to be a reasonable explanation as to why such a person would even be interested in or exposed to the topic to begin with?
"Physics IS mathematics" - more like physics is EXPLAINED with mathematics. I fixed that for you.
You don't need to know how to compute an integral or moreover know the details of the underpinning logic statement represented by an integral expression or how to prove one mathematically (not evaluate one) to appreciate the generalization that integrals measure areas, volumes, and higher dimensional analogies.
You don't need to know the chain rule, the general power rule, the product or quotient rule let alone the terribly complicated logical statements underpinning the notation to appreciate the generalization that derivatives are rates of change.
You don't need to know how to work with vector valued functions to understand that they can represent the objects and their attributes both macro and micro in a wide variety of circumstances.
You don't need to know how to express mathematics as a provable logical statements in order to appreciate that it works. Being able to evaluate an expression is not the same thing as proving the fundamental underpinnings that expression is talking about.
Physicists aren't generally sitting around proving new theorems (unless you're Witten, Werner, maybe a couple of others, or some of their grad students) and publishing them in math journals. Usually physicists are implementing an instance of some already established area of mathematics that seems to suit their theoretical needs so they can predict something and go look for it in a lab. Or they are just asking mathematicians for assistance like Weber and Gauss which is neither the first or last of such an academic coupling.
The point is most physics models are implementations of existing maths under the assumption that the math is logically sound and meaningfully applies to what they want to know. Such as using an integral to measure energy stored in capacitance. That's a task much more akin to computer programming than it is to the activity of mathematicians.
If someone can't adequately break down what they are doing to help someone who knows very little about the subject matter "get it" or at least appreciate its marvel on some level without following through the mechanics of some mathematical algorithm than I'd be worried that that person doesn't really know what they're doing and is resorting to "blue box" strategies.
Thought experiments (or summaries on scientific research and problems -- how do you expect to attract young scientists who aren't reading academic journals because they don't have access or are too young?) have been crucial to the development of the sciences so the outright dismissal of their merits because there is some sentiment that there should be some necessary technical threshold to be met before one can contribute shows a complete lack of understanding or appreciation of how much such discourse has progressed our understanding of mathematical and scientific principles. Even if it means addressing silly questions seriously sometimes.
So how does Wolfram's own creation Wolfram Alpha do in comparison against the other search giants?
Well they are not one in the same sure, but The Swedish Pirate Party also hosts The Pirate Bay itself so you can't completely separate them from each other either. The Pirate Party Becomes The Pirate Bay’s New Host
Obviously both sites and the Swedish Pirate Party are betting (pretty hard) on the election next month which a successful outcome would as previously posted put TPB and perhaps now wikileaks inside the Swedish Parliament.
Thanks for sharing I just grabbed the app from the Android store and it is pretty cool! :)
Now how to use what the phone can do to either scan for thermal or some other bio-signature and this thing is basically complete!
He has also seemingly updated the paper after receiving some feedback. This is the link to a version dated today: http://www.hpl.hp.com/personal/Vinay_Deolalikar/Papers/pnp_updated.pdf
Whoa its from the future! (article is dated June 26th)
That's some seriously scary stuff from these guys. Has EFF or any of the other groups issued a response to either of these pieces yet?
soundexchange -> broadcast (internet radio)
ascap -> performance
typed in too much of a hurry sorry. This stuff just really pisses me off about these kind of groups.
They probably are relying on or hoping to attain the same standing RIAA has gotten through SoundExchange to collect broadcast royalties even from non-members.
So there is clearly precedent that suggests ownership and membership are not sufficient concerns to these types of organizations. Unless it is their material or members that is!
So in this case they are either seeking statute authority to collect song composing royalties for members AND non-members, or they intend to behave that way anyway and defend it on the premise that the copyright office already delineated similar powers to SoundExchange and that since ASCAP is a similar group to SoundExchange they are entitled to a similar wide scope of authority (performance royalties -> SoundExchange vs composing royalties -> ASCAP)
I'd really like to see this blow up in their face and get both groups rights to even try this sort of thing revoked, but there are too many MAFIAA members in DoJ (and probably other parts of gov't) now and they have the administration's support (much to my dismay as I do generally otherwise support the administration). So this could get ugly and have bad consequences quickly.
I really hope the copyleft groups start gathering funds and resources in a way to respond to this head on. I'd support it.
About RIAA lawyers at DoJ:
http://www.wired.com/threatlevel/2009/04/obama-taps-fift/
About RIAA/SoundExchange:
http://www.dailykos.com/story/2007/4/24/141326/870
http://slashdot.org/articles/07/04/29/0335224.shtml
Numerous researchers in synthetic biology have been trying to do exactly that.
Here is an example Open Wetware BioBricks
Synthetic biology has two paradigms the first is the top down approach which deals with gene knock outs to look for minimal sets necessary for life that can then be tailored to suit specific needs/tasks. The other approach is the bottom up which have inspiration from the Miller-Urey experiments. They are trying to spontaneously generate a new biological system from scratch. Some researchers in this camp trying to create synthetic cellular components in hopes of putting all the synthetic parts together to create a functioning cell such as synthetic Golgi bodies
There has been some promising results from both approaches. It is a pretty exciting time in Biology.
I've seen many many posts talking about "you need logic but not really mathematics"
I'm not really sure what the confusion is about why mathematics would be separate from logic in any respect.
Symbolic Logic is Mathematical Logic. Principia Mathematica clearly makes the case that math and logic are the same thing. As did many works around the same time and since. So how can you need logic for computer science or programming yet somehow not mathematics it just doesn't make any sense to me.
As Dvorkin put it, it isn't as simple as always use classical methods or always use Bayesian.
I also think sometimes whenever possible you need to use both to help give you more evidence. Particularly if the results are somehow unclear. If they both give similar or approximately the same results that might be pretty good evidence that the final result is reasonable.
If you get wildly different conclusions then you need to consider why you got different answers and maybe it is because one method better models the situation than the other (and there are some tools to help figure that out in some cases) or perhaps the irregularity is identifying a problem with the data collected that would have gone undetected if you only used one or the other (and there are tools for when this happens too in some cases).
Very careful model building, very careful experimental design, very careful analysis, and very careful model adjustment isn't a guarantee that the results will be correct with any methods, but it is the best that can be done and is critical to always follow.
Like Daniel Dvorkin has said Devore's book Probability and Statistics for Engineering and the Sciences is an excellent starting point.
Definitely learn to use R since its free you don't have to worry about paying licensing fees. It is also widely used (no matter what you here from SAS, Minitab, SPSS, etc).
Books I would recommend that I think fit his other suggestions are Bowerman/O'Connell Linear Statistical Models: An Applied Approach and Wackerly et al Mathematical Statistics with Applications
Devore talks about Bayes Rule as does Wackerly and Wackerly's last chapter talks about some Bayesian techniques, but these are merely primers for what is typical in a Bayesian course. So I recommend these two books as analogous with Devore's: Bolstad Introduction to Bayesian Statistics and to Wackerly's: Hoff A First Course in Bayesian Statistical Methods
Some things you need from mathematics are the ability to integrate, work with matrices and matrix operations, and algebraic manipulation. Familiarity with transformations and operators especially linear ones is useful since many procedures in statistics are linear operators. The highest levels of statistics will get even more math intense using mathematical results from areas like ODE/PDE, Galios Theory, or general Measure Theory.
The wikipedia's statistics articles are pretty good overall, but as Dvorkin noted some are more technical than what would be friendly to those that are new to statistics. When you feel that's the case try using the sources linked as citations in the article or google confusing parts and it is generally possible to find an explanation for almost any background level.
However if you can get through these texts you're background would be pretty strong.
Science is in the business of probably knowledge. So they really need to improve their probability and statistics knowledge.
It is not a shortcoming of statistics that other people, like various scientists who aren't statisticians, don't know how to use or properly interpret statistics. It is a shortcoming of their knowledge.
It is not a shortcoming of the Copenhagen interpretation of quantum mechanics or the Chicago school of economics if I don't understand or know how to correctly interpret their results. It is my shortcoming and fault for not knowing enough to connect the dots.
I do statistical research some of that is through interacting with researchers in the biosciences. Often when I go to talk to a researcher and ask them if they could use some statistical or mathematical or computational assistance with their research it has almost always been a fruitful starting point to long conversations and getting into the research. Now sometimes it was simply a matter of looking at their F-test results or ANOVA scores and telling them what it meant (like with a regression model relating proportions of certain characteristics between taxa), more useful interactions for me often mean working on new algorithms or estimators or working with fitting a model from their empirical data because there isn't a reliable standard model to work off of (like intergenic distance between genes in an operon) that kind of challenge makes less engaging work worth the hassle. Maybe I'm odd because I've worked hard to have a good background in both statistics and biology, but I shouldn't be.
Although here is an observation that perhaps supports some of the intent of the article from my own experience. I was speaking with a biology graduate student and it came up that they had a biostatistics course in the department. Of course as a statistician my mind goes towards survival function, failure rate, life tables, censored data, bioassy, epidemiology, microarrays, clincal trials, topics along those lines. It turned out their course focused z tests, t tests, f tests, confidence intervals, point predictions, least squares regression, multiple regression, ANOVA, and things along these lines just with simulated problems in a lab setting. That is not necessarily a bad thing, but much of the core math was under played or missing like model assumptions and alternate formulations or things like dummy variables. The worst part was that even though they were doing well with the class they had no confidence in actually using the statistics and didn't understand how to interpret the meaning of something like a confidence interval, they knew how to calculate one, but it wasn't clear what it actually meant to them.
The corollary to the notion in the summary I'd rant and claim is that scientists overall have less than desirable skills in mathematics, statistics, and computation than those who studied those disciplines principally and that's hurting science. However many in those three disciplines really know little beyond basic results in any of the sciences which hurts the applicability of these mathematical fields to the sciences and likely hurt our ability to develop certain types of discipline specific results that can be generalized from work in application problems.
In either case whether you're a typical scientist or a typical math/stat/comp person in order to become proficient enough in the other areas it requires going an awfully long out of the way compared to any counterpart who simply does not care and goes straight through as many before have. While in some areas of research on either side it is no problem to do as has been done and not further knowledge into those other areas. Increasingly results that have the highest levels of impact are coming more and more from truly interdisciplinary research. In order to further encourage that for those who are interested in such fields (aside from making more clear what areas in any of the fields fringe to such interdisciplinary work) we need more incentive to study more than one field and/or better ways of enabling fruitful cooperation between the camps.
EMI and other Cope programs have MIDI output capability. He'd pipe the midi files to a program that allowed for printing as sheet music (I don't use apple's so I am not familiar with what program specifically he was sending the midi to for sheet music printing). The EMI described in Cope's books required manually keying in entries for the databases, but it would seem Zenph's only difference is that it is working with the audio capture directly. I'm pretty positive EMI or the other Cope programs could readily be adapted to do that. Also if EMI were only given a few pieces or a single piece in the analysis db and the parameters are set in a certain way EMI would do exactly what Zenph does.
I agree though that Zenph seems to want to head toward what EMI is already capable of. This gets compounded when you consider the style signatures or genetic analogies used by Cope or even a system like Pandora (which has been my impressions about what Recombinant might be all about) then it makes Zenph all the more trivial, but it points towards the increasing trend in this section of industry trying to do this sort of algorithmic composition, re-arrangement, remastering, etc.
I don't think that's a correct interpretation of copyright law.
"In the case of works made for hire, the employer and not the employee is considered to be the author." from LOC copyright circular
So if work for hire allows for corporations to create and author copyright materials then why wouldn't a corporation be able to author the copyright of the output of this sort of program?
David Cope's Experiments in Musical Intelligence and related works (SARA, other works, and his own company called Recombinant inc ) have been doing this for many years.
SIAM (Society for Industrial and Applied Mathematics) has several special interest groups related to computing/programming problems. The other major math and stat groups have excellent articles on computing problems from time to time as well like the AMS, MAA, or AMSTAT, but SIAM probably provides the most of these groups and a lot of coverage that compliments IEEE and ACM. Also depending on if you're working in a specific industry or if you're furthering your studies in a graduate program there may be other professional societies that deal with informatics or computational issues related to that focus.