The mathematics of discovering new things
washingtonpost.com
The mathematics of discovering new things
1–10 of 10 posts
Re: The mathematics of discovering new things
#2I love how after I cofounded a math game company, there seems to be more and more math everywhere I look. "Blue car syndrome" at its finest!
For me this article especially rings true with self identity / confidence. Since four years ago I started learning how to program video games, I recognized that internally, "Well if I could totally change my career overnight, and open up a plethora of new possibilities, then I wonder how many additional choices I will now start to see over time?"
I saw it as a graph of my life possibilities, with the first 25 years of my life bumping along in the 1-10 range, then a sudden inflection point to 100. Looking at this graph objectively, one can imagine a second inflection point that goes to 1000, etc.. maybe leading to an asymptote! But I digress.
Re: The mathematics of discovering new things
#3I love how after I cofounded a math game company, there seems to be more and more math everywhere I look. "Blue car syndrome" at its finest! For me this article especially rings true with self identity / confidence. Since four years ago I started learning how to program video games, I recognized that internally, "Well if I could totally change my career overnight, and open up a plethora of new possibilities, then I w…
> "Blue car syndrome"
There's a Name for That: The Baader-Meinhof Phenomenon[0], also known as the "Frequency Illusion."[1]
[0] http://www.damninteresting.com/the-baader-meinhof-phenomenon...
[1] http://en.wikipedia.org/wiki/Baader-Meinhof_phenomenon#Frequ...
Re: The mathematics of discovering new things
#4So if I get it right:
Say you have elementary concepts A, B and C. Combinations of these concepts also give something, e.g. AB, AC, etc. You have tried and exhausted all possible (and sensible) duplets, triples, etc. and now you are bored.
Then something new comes along, D. Now you not only have a new elementary concept, but a whole lot of new possible combinations with previously known elementary concepts: AD, BD, etc.
Re: The mathematics of discovering new things
#5You can click through the full Nature article, which is nice. I haven't read the article in depth, but it looks like interesting stuff. A bunch of models in Bayesian non-parametrics are, like the described model, based on Polya urns. It should be a simple matter to use this model for some interesting clustering tasks, perhaps as an alternative to the CRP.
Re: The mathematics of discovering new things
#6I love how after I cofounded a math game company, there seems to be more and more math everywhere I look. "Blue car syndrome" at its finest! For me this article especially rings true with self identity / confidence. Since four years ago I started learning how to program video games, I recognized that internally, "Well if I could totally change my career overnight, and open up a plethora of new possibilities, then I w…
> Looking at this graph objectively, one can imagine a second inflection point that goes to 1000, etc.. maybe leading to an asymptote!
It looks more like exponential growth to me! Or are you hoping to asymptotically approach death? :)
Re: The mathematics of discovering new things
#7I love how after I cofounded a math game company, there seems to be more and more math everywhere I look. "Blue car syndrome" at its finest! For me this article especially rings true with self identity / confidence. Since four years ago I started learning how to program video games, I recognized that internally, "Well if I could totally change my career overnight, and open up a plethora of new possibilities, then I w…
> "Blue car syndrome" There's a Name for That: The Baader-Meinhof Phenomenon[0], also known as the "Frequency Illusion."[1] [0] http://www.damninteresting.com/the-baader-meinhof-phenomenon... [1] http://en.wikipedia.org/wiki/Baader-Meinhof_phenomenon#Frequ...
"Blue car syndrome" is also a name for that, and an older one than "B-M".
Re: The mathematics of discovering new things
#8Huh? A common representation of the Polya urn which results in a Dirichlet process corresponds to "novel" colors as well. If you pick a normal color you put it back with another ball of that color, but if you pick the black ball, you can pick an entire new color, hence discovery of a new thing. You can read it on Wikipedia, the source of all knowledge. :-) https://en.wikipedia.org/wiki/Dirichlet_process
Original link is by the way: http://www.nature.com/srep/2014/140731/srep05890/full/srep05...
Shameless plug: if anyone wants to work with me on application of nonparametric Bayesian methods to 3D navigation and scene reconstruction, please do so. https://github.com/mrquincle/octave-scripts/tree/master/thes....
Re: The mathematics of discovering new things
#9So if I get it right: Say you have elementary concepts A, B and C. Combinations of these concepts also give something, e.g. AB, AC, etc. You have tried and exhausted all possible (and sensible) duplets, triples, etc. and now you are bored. Then something new comes along, D. Now you not only have a new elementary concept, but a whole lot of new possible combinations with previously known elementary concepts: AD, BD, e…
I've observed this with fandom crossovers.
Re: The mathematics of discovering new things
#10Very interesting. Is there a corollary to this? Say, if a person or a bunch of people are curating articles then they will select only those which are most novel and all other related things will get dropped. In that sense, we should not see similar stuff on the front page of HN. Yet, I recall, seeing at least 3 news items about Microsoft retrencing people, on the front page of HN. So should HN consider applying the corollary "correlated novelty".