Created
February 2, 2018 00:06
-
-
Save krishnanraman/324206b7b4eb7ede6a532093ef8f1282 to your computer and use it in GitHub Desktop.
ideas.txt
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
distribution data => distribution name, params supervised ANN classifier | |
eg. | |
Poisson(5) data 1000 samples => ("poisson", 5) | |
Uniform(10) data 1000 samples => ("uniform", 10) | |
Normal(mu, sig) data 1000 samples => ("normal", mu, sig) | |
etc. | |
synthetic data - https://arxiv.org/abs/1804.06516
make lots of synthetic scenes train on them and use model on test. placing cars in arbitrary scenes. very similar to generating swish by randomly iterating on diff activations until one happens to work. generate millions of distributions by varying mu sigma until you learn gaussian ....
detection of workout videos based on content ( rhythmic beat (audio FFT), racy content, text tags by actor actress lyricist md etc )
simulate non-binom ( gaussian poisson etc) with coin toss
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
classify humans beyond endo ecto meso.
based on gross features ( wt height chest waist measurement) and phenotype from dna. fit every human into a class. each class has ideal mu sigma weight. ppl can migrate between classes due to change in gross.