-
Dear authors,
Thank you for sharing the official implementation of your work. I'm particularly interested in reproducing the adversarial training results using your Balanced Adversarial Focal Train…
-
https://github.com/MadryLab/robustness/blob/a9541241defd9972e9334bfcdb804f6aefe24dc7/robustness/attacker.py#L195
Assuming that you are solving for per-datum perturbations, and not a broadcasted (or…
rsokl updated
2 years ago
-
Beta parameters define the perturbation kernel of the SDE. This in turn effects the size of the perturbation of the original data as a function of the time parameter.
There are two parameters, beta m…
-
the feature perturbation and vae perturbation, which one is more useful
-
Right now, separate function for perturb_spar, perturb_hnpi, etc, can all be combined into a single file to avoid redundancy and to make it easier to edit them.
Additionally, these should be updated…
-
Hi, I ran the scripts provided in this github repo but faced multiple issues:
**Issue 1: Cannot reproduce the paper results**
In Table 4 of the paper, it is mentioned that the word level perturbat…
-
Advantages:
1. Probably the easiest and quickest to setup
2. Realistic evolution
Disadvantages:
1. Lots going on apart from the NAO,
2. No control simulation to remove drift/other sources of variabi…
-
A program was [posted to golang-nuts](https://groups.google.com/forum/#!topic/golang-nuts/U7plctxPET4), with the question “is this code thread [safe]?”
The program has a race: it executes a `select…
-
-
I'm having troubles downloading the LINCS L1000 perturbation signatures:
```{r}
> l100.genetic.pertsig l100.compounds.pertsig