The Magic Café
Username:
Password:
[ Lost Password ]
  [ Forgot Username ]
The Magic Cafe Forum Index » » The Science of Magic » » PHD ideas for NYC University (0 Likes) Printer Friendly Version

Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
I love magic and mentalism. I have a psychology and computer science undergraduate and computer science MS degrees and am thinking about doing a PHD. I want to find a PHD topic/field that is magic related. Very little I can think of in Computer Science – but if someone CAN please post here. I realize Psychology has more potential, for example I recently read Sleights of Mind and the authors seem to be researching attention etc. Please post any interesting research ideas specially one you think or know some university in NYC would take seriously.
Brad Burt
View Profile
Inner circle
2675 Posts

Profile of Brad Burt
The psychology of using computers for deception: How are we fooled by trusting in computers?

The psychology of 'magic' as 'puzzle' and how/if/whether there is a positive effect on human neurology?

Meta-Ethics and deception in computer use and magic: Metaphors for a Post-Luddite world

Belief in 'magic' and belief in 'computers': A substitute for belief in God. (Granted a secular university may not be jiggy with this one, but I still think it's interesting and I'm just brain storming anyway.)

The computer and it's use as a auxillary theraputic device.

Well, if I think of anything else I'll pass it on. My wife has Sci.D. in Psychology by the way, so I'm familiar with all these disciplines. Would be interested in what you come up with. If you want to bounce anything off me feel free.

Best,
Brad Burt
Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
Thanks Brad.
If I went psychology route, I too have some ideas that are interesting, however , ideally , I was trying to think up a more quantitative computer science thesis - yet relating to magic in some way. This might be too difficult to do, unless I am going to go for something extremely contrived. For example, deck of cards can be viewed as a set of 52 different objects, which can be linked to a concept of a static array in computer science and one can study some abstract operations on a small array that are in turn linked to what you can do with a deck of cards, but its extremely contrived.

Something a bit less quantitative but cool I was thinking of is , as computers get more intelligent, should or shouldn't THEY get fooled by magic? Magic Turing test of a sort for AI.

All your ideas are interesting. Out of the bunch, I like the 'positive effect' question the most. Its one of those topics that is "neat" and people outside psychology *might* be interested in. Also, it sounds like research would require showing tricks to people which is an awesome element. Long, long time ago I wrote a paper of effects of puzzle video games on child development - this would be related (and there is a ton of research in that area).

Post-Luddite world would be a great place to live in Smile.

Did your wife get the degree in NYC ? If so, I am guessing from Pace university? Its the only one around that has professional doctorates.

Thanks again for your response.
Yellowcustard
View Profile
Inner circle
New Zealand
1312 Posts

Profile of Yellowcustard
A group of pepole from queen Marys collage london looked in to how the matmatics behide card tricks is the same that is used in comptuer science. They also say that the maths in magic has influced some recent devlopments in comptuing. There research is quite involed yet they also publish this little booklet as well. I like it as its a way in for any one and they will learn a trick and understand a bit of comptuer science.

I don't know if its what you want but check it out.

Heres a link to a down load PDf http://www.cs4fn.org/magic/downloads/cs4fnmagicbook1.pdf
Enjoy your magic,

and let others enjoy it as well!
jfquackenbush
View Profile
Special user
Out here on the desert
607 Posts

Profile of jfquackenbush
Read Hubert Dreyfuss's book "what computers still can't do." It's a phenomenological critique of trends in AI that were prevalent at one point, but there's an interesting application there that I would love to see some PhD work on. I think one can make a case that being susceptible to deception is a necessary element of human intelligence, and therefore a true "thinking machine" would need to have the same susceptibility. I think an AI that could make useful application of the "errors" of human reason, such as confirmation bias or some of the game theoretic problems that we're naturally bad at (like the prisoner's dilemma) would be really interesting. Conceivably, these faults in our thinking have some sort of adaptive advantage from an evolutionary standpoint, and while I'm not up to date on the field and don't follow it closely, I think trying to locate and apply those advantages in AI research might open some interesting avenues for advanced Comp Sci work.
Mr. Quackenbush believes that there is no such thing as a good magic trick.
Brad Burt
View Profile
Inner circle
2675 Posts

Profile of Brad Burt
The Metaphysics of Computers: How Computers are the Arch-type of "The Magician" in Jungian Theory (Crap, qualitative again....)

How Computers are Like a Magician: A Quantatitative Analysis of the Mathmatical Permutations Inherent in Magic Misdirection

I have absolutely NO idea what that last means, but it sounded so cool that I just had to toss it out there.

The questions that occur to me off hand are:

How is a computer 'like' a magician or like a magic trick?

How could computers be used to enhance magic as a performing craft?

If a computer 'could' do magic what would that magic consist of, look like, etc.?

How would you 'quantify' anything? Simple description of "just what it is and consists of?"

It seems to me that putting computer science together with magic you have only a couple of avenues:

#1- How computers could be used in calculating the math of certain types of card tricks

#2- How computers could be used to organize a magic act

#3- How computers could/are used to DO magic effects of various kinds and why

Hope this helps at all, good luck,
Brad Burt
Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
I had though about that angle a bit jfquackenbush,and I know I will get off topic at hand but , I am not so sure being susceptible to deception IS a necessary element of generic intelligence(nothing special about human). I think deception is a side effect of something else useful or what used to be useful in our evolutionary past, but frankly , as sad as it would be for me as a magician, if people didn't get deceived easily or at all, the world would be better for it. Same goes for all errors in human reason. I think in the future, if we merge with computer technology more tightly some of these errors will become obsolete.

Yellowcustard you gave me a lead to an interesting book by S. Brent Morris who studied shuffles for his PHD.
Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
Quote:


#1- How computers could be used in calculating the math of certain types of card tricks



This is a direction that people had taken before. S. Brent Morris who's work I just learned about has a book on faro shuffles. I have to think about this line some.
jfquackenbush
View Profile
Special user
Out here on the desert
607 Posts

Profile of jfquackenbush
To clarify, It may not be a necessary element of generic intelligence, but it might be to human intelligence. I think the question that might be worth asking is how do inductive leaps and errors of logic serve human intelligence and how might those blind spots serve AI. That said, I'm skeptical that pure reason implemented in a thinking body is possible or even particularly desirable, but that's more of a philosophical than a scientific debate. Best of luck to you and please do keep us abreast if you decide to pursue it.
Mr. Quackenbush believes that there is no such thing as a good magic trick.
Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
Jfquackenbush I studied a little AI in grad school, I am not sure that's where I want to go but in case you are interested, most complex problems are not solved completely but rather approximately, using different methods / heuristics - they will by definition lead to blind spots of different kinds ( this is how human blind spots came about too I'd wager.) When you beat a computer at GO, you 'fooled it' .
jfquackenbush
View Profile
Special user
Out here on the desert
607 Posts

Profile of jfquackenbush
I get what you're saying, so let me just clarify once again. I thought one of the things that was most interesting about watson's run on jeopardy was the way that it made a mistake that no fairly well-educated person would be likely to make by answering Toronto for a question about a US city. The thing that strikes me as revealing about that kind of error is that it indicates the sorts of associations and probabilities that humans take into account when we're working off of incomplete information. Reading the explanation for why watson made that error was very revealing from a non-specialists standpoint about how Watson makes decisions, and I'm curious if a similar approach might reveal something about human intelligence. What strikes me as interesting about the question of susceptibility to deception is, does that come as a product of how we make intuitive leaps, and can the particular way we're susceptible to deception reveal something about how we make those leaps as opposed to how AI implementations would work. For example, presumably it isn't difficult to program a computer to avoid the simple errors in inductive reasoning that results in the phenomenon of confirmation bias in humans. What I'm curious about is how does that process work in humans such that we are susceptible to confirmation bias, and are there advantages of that process over another possible process such as might be implemented digitally that wouldn't be susceptible to that sort of error.

Anyway, this totally isn't my field, but I'm very interested in the discussion and I really do hope to speak to you about your work more in the future.

All the best,
J
Mr. Quackenbush believes that there is no such thing as a good magic trick.
Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
I agree its an interesting topic of conversation. It might be less fun tying this to a quantitative research of some kind.
I can think if many examples of how confirmation bias could have been useful heuristic in our past (but , just my opinion, it outlived its usefulness). I think being aware of it makes you a bit less prone to fall victim to it.

Obviously, I find all cases when our reasoning/intuition is wrong very interesting (magic exploits many). I was recently thinking about the famous Monty Hall problem - why would we evolve statistical misunderstandings like that?
jfquackenbush
View Profile
Special user
Out here on the desert
607 Posts

Profile of jfquackenbush
The Monty Hall problem is a good one, and one that even if you're aware of it's hard to avoid. In fact, I think the way people deal with probabilities in general is strangely poor. I used to work in a job where I did statistical analysis of survey responses for large companies, and they were always asking me questions that the sample sizes were insufficient to answer with any confidence. It felt like a a conversation I was forever having, explaining and reexplaining error margins, and in the meantime less scrupulous people were giving them answers and they were making million dollar decisions based on results with error margins as high 60% in some cases. It was truly baffling until I realized that what people really wanted was cover for decisions that they were making for other reasons. In a way they were just using statistics as a form of misdirection, although they probably didn't even realize themselves that's what they were doing.
Mr. Quackenbush believes that there is no such thing as a good magic trick.
Damon Zale
View Profile
Regular user
180 Posts

Profile of Damon Zale
I hear ya J. In fact, my pet peeve is almost the same thing, anecdotal evidence. 'My friends grandfather smoked all his life, ate triple hamburgers, drove a motorcycle, had unprotected sex ... and lived to be 90' type of statements. BTW I wonder if it might be very bad for some people’s health to have such acquaintances when combined with lack of statistical understanding? This can make a correlational study Smile.
It seems I come across it all the time [but then maybe I just notice it over people being rational ala confirmation bias Smile ].
stevemcd
View Profile
New user
1 Post

Profile of stevemcd
I may be a bit late but here goes:

If you are interested in the algorithmic / applied stats side of card shuffling you may like to check out the work of Persi Diaconis ( http://www-stat.stanford.edu/~cgates/PERSI/year.html ). In particular:

1. Trailing the Dovetail Shuffle to its Lair. D. Bayer and P. Diaconis, Ann. Appl. Prob., 2(2):294-313.

2. Magical Mathematics: The Mathematical Ideas that Animate Great Magic Tricks by Persi Diaocnis & Ronald Graham, 2011. Princeton University Press.

Diaconis et al. ask questions such as 'How many times must a deck of cards be shuffled for the deck to be in close to random order?' one might generalize this to 'How many times must an iterative procedure be run?'. They provide asymptotic results for interesting cases (e.g. deck of 52 cards). Further work can be found on his homepage.

Disclaimer: I have not had a look at [2] yet but it sounds interesting / relevant Smile


This line of research, mathematical models of shuffling might be traced back to information-theoretic work carried out at Bell Labs by Gilbert and Shannon in the fifties.

In a slightly different area:

My own PhD work is currently related to machine learning and computer vision. An interesting idea in this area (learning-based video tracking/gesture recognition project) would involve training a classifier to 'learn' and be able to recognise subtle differences of human motion, captured by a (high frame rate?) camera. For example how about training a (machine-learning) system on a three card monte game? this might involve showing it many examples of a throw and many examples of a h*pe or say a single and d**ble lift. Can we then extract meaningful features from these data that separate them in some subspace? Are we then able to accurately classify a novel (previously unseen) instance correctly? in real time? can we generalize between different people performing the same move? This sounds similar in nature to previous sign language recognition work.

You might however have difficulty in convincing people to let you record live monte test data Smile

HTH,
steve
The Magic Cafe Forum Index » » The Science of Magic » » PHD ideas for NYC University (0 Likes)
[ Top of Page ]
All content & postings Copyright © 2001-2022 Steve Brooks. All Rights Reserved.
This page was created in 0.14 seconds requiring 5 database queries.
The views and comments expressed on The Magic Café
are not necessarily those of The Magic Café, Steve Brooks, or Steve Brooks Magic.
> Privacy Statement <

ROTFL Billions and billions served! ROTFL