private:running_notes

This shows you the differences between two versions of the page.

Both sides previous revision Previous revision | |||

private:running_notes [2019/12/27 21:27] admin |
private:running_notes [2019/12/27 21:28] (current) admin |
||
---|---|---|---|

Line 9: | Line 9: | ||

* Fast dropout (or any dropout) should zero mean the input - that way dropout doesn't change the weights. | * Fast dropout (or any dropout) should zero mean the input - that way dropout doesn't change the weights. | ||

* Values GAN (or classifier) - pain/pleasure classifier on latent space, use to guide the search when trying to solve a problem. | * Values GAN (or classifier) - pain/pleasure classifier on latent space, use to guide the search when trying to solve a problem. | ||

- | * Take a space, second order decorrelate, model with BIC diagonal GMMs, expand as c_i * x_j, truncate those that are small, repeat until no signal left. Inverse is to generate from latent/noise model, apply inverse to get c_i * x_j, estiamte c_i, fill in missing from noise distribution, sum to get x_j and repeat downwards over. | + | * Take a space, second order decorrelate, model with BIC diagonal GMMs, expand as c_i * x_j, truncate those that are small (as a fraction of signal power?), repeat until no signal left. Inverse is to generate from latent/noise model, apply inverse to get c_i * x_j, estiamte c_i, fill in missing from noise distribution, sum to get x_j and repeat downwards over. |

private/running_notes.txt ยท Last modified: 2019/12/27 21:28 by admin