Does anyone know what would be a good source of bibliography for this talk? I'm more interested in the math aspects of what he is doing, for instance, where the topology comes to the play.
Oh that's great! (You then may know more math than I do, haha). You might want to start with this introductory/programmatic paper by Carlsson himself: "Topology and Data" www.ams.org/journals/bull/2009-46-02/S0273-0979-09-01249-X/S0273-0979-09-01249-X.pdf If you're unsure about some construction or concept you may consult the references therein :) Hope that helps
Actually here he describes Mapper tool and there is no topology at all. But if you are interested in topological methods, there is another tool, widely promoted by Gunnar Carlsson called Persistent Homology geometry.stanford.edu/papers/zc-cph-05/zc-cph-05.pdf
That was absolutely bizarre at the end. Anyway, does anyone know if we can use any of these methods to preprocess data and pass it, to say, a neural network?
![ILLSLICK - KILLSHOT REMIX [FULL VERSION]](http://i.ytimg.com/vi/RIK8RrqIAzE/mqdefault.jpg)
I'm only 10 mins in, and I've never been so into a educational video! So fluent! so clear!
Great talk with an absolutely bizarre ending. I hope security was keeping a close eye on that guy as he hugged Gunnar. So weird.
Gunnar you are superb ..the way u put down machine learning field is worth admiring
Does anyone know what would be a good source of bibliography for this talk? I'm more interested in the math aspects of what he is doing, for instance, where the topology comes to the play.
What's your math background?
@@janouglaeser8049 i've graduated in pure math.
Oh that's great! (You then may know more math than I do, haha).
You might want to start with this introductory/programmatic paper by Carlsson himself:
"Topology and Data"
www.ams.org/journals/bull/2009-46-02/S0273-0979-09-01249-X/S0273-0979-09-01249-X.pdf
If you're unsure about some construction or concept you may consult the references therein :)
Hope that helps
Por cierto, ¿hablás español?
Actually here he describes Mapper tool and there is no topology at all. But if you are interested in topological methods, there is another tool, widely promoted by Gunnar Carlsson called Persistent Homology geometry.stanford.edu/papers/zc-cph-05/zc-cph-05.pdf
That was absolutely bizarre at the end. Anyway, does anyone know if we can use any of these methods to preprocess data and pass it, to say, a neural network?
You definitely can if it make sense for your task
He himself bridges the gap in th-cam.com/video/i-tpAxnrQ8s/w-d-xo.html
thanks for the video presentation!
Thank you for the talk, very clear
Very interesting talk. Was wondering, wouldn't the choice of a particular data projection, e.g. density estimation, affect the final graph?
Yes, any metric used for data projection results in a different final topology.
Very illuminating talk.
1:00
10:44
12:56
How much computer programming is required for doing TDA ?? I'm not good at programming . Will it gonna be really difficult for me do TDA ?
Yes
You can pick up on this stuff still. Just use python like most people do
to implement it no, to understand it yes