haliax

While there is nothing inherently wrong with making a profit from data about users, there is quite a bit wrong with the way it's being done on Facebook, and much of the rest of the web/mobile for that matter.

For one thing, services fail to properly inform their users of what they're doing. It's quite common for ordinary users to either not know that their data is being collected and used in those ways, or the extent of the data which is being collected (e.g. Location, Phone Identifiers, and sites visited across the web).

For another, opting out is generally designed to be quite difficult. Unless you're vigilant enough to know which hosts to block (among other things) it's almost impossible to keep your browsing history out of the hands of the hands of third parties (especially with all the different ways browsers can store information, see, http://samy.pl/evercookie/). This is the case even if you do not sign up for or use a service like Facebook.

Also, as someone noted below, there is an issue with the simple fact of the data's existence, since from that point on it is out of my control, and can be used by anyone who can get their hands on it, in any way, which may well be harmful to me.

> I think ANN is the real thing, and most of the other "AI" solutions that we have give the impression of intelligence, but are simply algorithms that solve problems we have in very well defined ways...ANN is a general intelligence that gets shaped by experience and adapts figure out its own solutions...It isn't some coder banging on the keyboard trying to replicate what a human would do under a specific situation, it is something that truly learns.

It's not ANNs on the one hand and "some coder banging on the keyboard trying to replicate what a human would do" on the other.

Supervised learning is the most common use case for an ANN, and means learning from labelled training data during a training phase and then not changing; it can be done with a number of techniques, including Support Vector Machines, Decision Trees, Bayesian methods, and so on.

Reinforcement Learning on the other hand, is continuous learning from trying things and making mistakes -- experience, in other words. It too can be done a number of ways, like with Evolutionary Algorithms, Markov Decision Processes, Inductive Logic Programming, etc. Hell, PG's spam classifier learns from experience, and it's nothing more than a Naive Bayes classifier ( http://www.paulgraham.com/spam.html ).

Without going on for much longer, my point is that there really is no reason to exalt ANNs the way you have.

Disclaimer: I'm a huge believer in Artificial General Intelligence via Competent Program Search ( see http://metacog.org/research.html + http://www.google.com/url?sa=t&source=web&cd=2&v... + www.idsia.ch/~juergen/ )

> I guess this depends on how advanced your toolkit is.

I'm assuming you mean this to counter the "simplest of biological neural networks" part, rather than the "completely opaque black boxes" part. Nonetheless, I think it is fair to say that ANNs still fall far short of the complexity of even the simplest biological brains. For example, even the nervous system of a Hydra features neurotransmitters [1] (as opposed to just heterogeneous signals of ANNs) and brains are affected by their own electric fields [2].

> You want to be able to discern some rule base from the system that you can understand? Even biological brains do not have that property.

More generally, transparency is important to be able to understand the inductive bias you've built up, and to try to alter or refine it in useful ways.

Also, the fact that our brains lack transparency doesn't justify leaving it out of an AI system, nor does it demonstrate the difficulty of building a transparent AI -- nature just had no drive towards transparency. Plus we (humans) can introspect and explain our reasoning.

> See, but the problem is that we don't even have the knowledge of how real brains work to start forming such principals > based on a technique which reflects the structure of the problem of learning, generalization, and hypothesis search and illuminates it.

This is only the case if we're speaking about biological brains, and not as a generic word to mean "intelligent system". In the latter case, we do in fact, have quite a bit of knowledge about such principles from reasoning about hypothesis spaces. From where we get things like active learning, or Solomonoff's work on Universal Induction [3].

> It's not going to be a fully general, "human-level" AI, obviously, as that requires huge amounts of semantic knowledge about the world that was encoded through billions of years of evolution

By "semantic knowledge" do you mean inductive bias? Because otherwise I'm at a loss. I don't believe that the picture for Artificial General Intelligence is as bleak as you make it sound though.

As an aside, you mention "Lots of more biologically-inspired approaches have been explored". Do you know of any projects looking to mine the structure of various parts of brains to figure out the sorts of inductive bias those structures correspond to? (As opposed to just copying structure) What I mean, is that presumably if a part of the brain heavily involved in recognizing faces has a unique structure/wiring, that structure is optimized such that it performs well on face recognition -- and correspondingly poorly on something else, per no free lunch -- and that optimization should tell us something about the nature of recognizing faces. Sort of in the same way that the use of a Naive Bayesian classifier rather than a Bayesian Net might tell you that the classifier is optimized for cases where the variables are independent.

[1] http://books.google.com/books?id=WWN_t498S5IC&pg=PA12...

[2] http://www.scientificamerican.com/article.cfm?id=brain-elect...

[3] http://world.std.com/~rjs/pubs.html

Completely disagree about Ruby on Rails, the experience of using happstack ( http://happstack.com/index.html ) blows it out of the water.

The RAM Cloud database is really natural to use; it eliminates problems like object-relational mismatch and SQL injection by letting you do everything in idiomatic Haskell.

Plus, the type-system, monads, and algebraic datatypes make it much easier to write clean, correct code. Albeit, only once you've passed the learning curve.

Is there any particular reason I shouldn't want to send advertising emails, generate pages, or buy domain names en-masse with the hope of generating revenue? I'm not doing anything harmful (hence the exclusion of phishing) -- and I'm making money.

People value all sorts of ridiculous things like homeopathic "remedies", psychic readings, and bullshit seeds on Farmville, how are those any better than spamming or content farming? I'd argue that the first two, at least, are far worse.

For that matter, how is the threshold of "acceptable value" set so that we don't similarly condemn, Facebook or 37Signals or SEO firms, for doing work of much less value than say, drug discovery or computer security -- especially when these are all within reach of the same CS grads?

Do Google's adwords count as creating value? Because they're how it makes it's money, search -- which creates value -- is just used to bring people in. If that separation of value creation/profit is all right, then is it okay to make millions spamming and give it to an efficient charity (as below)? If not, why not?

Nice hack, and a cool way to get your feet wet with developing Android apps as a business.

A find analogue would be a cool extension of this idea: it'd be cross-language tool that shows you the definition of a given function, or where a function is used, or otherwise lets you run simple queries on your code base.

This seems to be a related idea: http://people.csail.mit.edu/rivest/Chaffing.txt

I graduated from the Comlab / New College last year. Which college are you at? What year?

Why couldn't this co-exist with a hirerarchical filesystem, maybe via a special subdirectory as procfs does? (With homonym files getting some distinguishing prefix, possibly based on their hierarchical paths.) It seems somewhat like a more sophisticated version of spotlight in that it can handle logical relationships between tags.

> I'm not sure anyone is arguing against the latter part of your statement but you as you contradict yourself by arguing that it's a software problem first. AI is a biological, psychological, cog sci, EE/hardware, et al problem because, as you state, understanding how general intelligence works is important--as it is important to know how to build the technology that will run it (ie, possibly the memristor + something else, of course).

AI is a software (or other implementation of a specific algorithm) problem first, because general intelligence boils down to having an algorithm along with the right sort of inductive bias, and then running that in a world where it has input and its outputs matter.

While the other fields you mention may have something useful to say about what that inductive bias is, that doesn't mean that the core problem isn't one of software. The parent's point, as I read it, is that the specific type of general purpose hardware you run it on is irrelevant to solving that problem, and thus that this article is really rather sensational.

Apart from that, since we don't know what algorithms we need to solve that problem, it seems like premature optimization to start coming up with new hardware configurations for it. Although I do agree that parallelism seems to be important (e.g. for evolutionary methods) there's no reason that something standard like a GPU can't do the trick as well as it does for other types of computations that need parallelism, or at least prove the concept.

I imagine the DARPA grant is for something more sensible though, like for being able to run known algorithms (or adaptations thereof) which already have known uses, much, much more quickly; or at greater scale.

If the parent means the No Free Lunch Theorem, then the point is mistaken. That theorem says that you can't improve the performance of a classifier for some objective function, without making it worse on another -- in other words, all algorithms have an identical mean performance when averaged over all possible objective functions.

The reason this doesn't mean that human-level AI is impossible is that we too are designed (well, evolved by natural selection) to perform well for a particular objective function: one in which say, the standard laws of physics/optics apply. Optical illusions illustrate that our performance on this objective function is not perfect.

Moreover, you can see a human being's performance on a different objective function by, for example, trying to recognize objects in pictures which have been scrambled according to some predefined method (e.g. shuffle the pixels but use the same random seed each time). Each scene will still convey the same amount of information about the objects in it, but it'll be pretty tricky to recognize the objects.

I was at New

It's doable at Oxford as well, though not terribly common.

I'm an American who applied and got in 4 years ago (graduated with an M.Sc. this past October).

The application process is phenomenally simple, especially when compared to that for American universities.

1. You fill out an application for Oxford, containing basic biographical information and some information about the course and college you want to apply for.

2. You fill out a more detailed UCAS application, with your academic history, etc. and write a personal statement

3. You get called for interview either by phone, in person in New York/Vancouver, or in person in Oxford depending on what you can do.

4. You take an exam in your subject of choice.

All of this happens between October and November, and you get a response in late December.

Your choice of College is somewhat important, as it determines the life you'll have at Oxford, so research that in some depth. I went to New, and loved it, and my thesis supervisor was at Magdalen which is just lovely in every conceivable way (he's also a pretty kickass Computer Scientist).

Academically, you'll want a decent SAT score, a good mix of SAT II's and AP's, and a good recommendation from a teacher in a relevant subject area.

Which College were you at? I went to New!

I did much the same, also, Google says you were at Wadham College, right next door to me!

Unlimited vacations? Can you elaborate?

Do you know of any good texts/articles/case studies on the subject? This sounds pretty interesting.

Mind if I pick your brain about this? My aim is thelemonkiller and my email address is devan.lochees@gmail.com

Masters in Computer Science, experience in Program Analysis and Machine Learning. Experienced with C, Python, Perl, Java.

Interested in lots of things, including: synthetic biology/bioinformatics, finance, cryptography, natural language processing, evolutionary algorithms, and quantum computing.

East or West Coast US (or Chicago) is cool with me, as is telecommuting.

contact devan.lochees at the mail service provided by gewgl

Is this true of all cravings? I mean, there are loads of times (now) when I crave Sour Patch Kids or something like that, and I feel like I am just after the taste.

Wow! First off, thank you for all of the very well informed responses!

I had no idea there was so much to be known about eating better, and why we eat the way we do, especially with regards to carbohydrates.

I should probably add that it's not just bad carbs I'm out to cut. I want stuff like steak/bacon/pepperoni pizza/nachos/taco bell all the time, and I know that with my genes it's only a matter of time before that sends me to the hospital.

Also, I'm quite skinny at the moment, apparently regardless of what I eat...so I'm not sure if that's a sign of a contributing factor

Again, Thank you so much for the information!

Why is that?

PERFECT! Thanks so much for the link to Raphael!

Do you know of anything "wow" inducing in Scheme?

I too started with Basic, and learnt Python and so on, but I feel like she likes the visual aspect of it and the link to HTML which she "knows" already

I'm planning on getting her that if it seems like a lasting interest and not just a passing one :)

Is he an academic and are there published papers about this? Actually, does anyone know of good (broad/accessible) papers on behavior in general?