We lifted 8 functions out of this paper's own repositories and ran 7 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| lsgos/uncertainty-adversarial-paper | canonical | 7 of 8 |
| Function | Status | Where it lives |
|---|---|---|
| H | Ran | lsgos/uncertainty-adversarial-paper/latent_plots.py code served (permissive licence) · get_code("5cd82d77453fe113") |
| crop_center_or_reshape | Ran | lsgos/uncertainty-adversarial-paper/src/utilities.py code served (permissive licence) · get_code("30af533f70831f9f") |
| define_VAE | Ran | lsgos/uncertainty-adversarial-paper/train_mnist_vae.py code served (permissive licence) · get_code("8b1b6b0189468dfd") |
| get_uncertainty_samples | Ran | lsgos/uncertainty-adversarial-paper/latent_plots.py code served (permissive licence) · get_code("ff1b724e6343f9ed") |
| imagenet_deprocess | Ran | lsgos/uncertainty-adversarial-paper/src/utilities.py code served (permissive licence) · get_code("2d51388b7dd28bbd") |
| make_grid | Ran | lsgos/uncertainty-adversarial-paper/src/utilities.py code served (permissive licence) · get_code("e2c8245dbaa9881e") |
| make_random_targets | Ran | lsgos/uncertainty-adversarial-paper/ROC_curves_cats.py code served (permissive licence) · get_code("541c8add17d0bd10") |
| create_interpolation_dataset | Not yet run | lsgos/uncertainty-adversarial-paper/interpolation_comparison.py code served (permissive licence) · get_code("7ae0ad52e475b4da") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Measuring uncertainty is a promising technique for detecting adversarial examples, crafted inputs on which the model predicts an incorrect class with high confidence. But many measures of uncertainty exist, including predictive en- tropy and mutual information, each capturing different types of uncertainty. We study these measures, and shed light on why mutual information seems to be effective at the task of adversarial example detection. We highlight failure modes for MC dropout, a widely used approach for estimating uncertainty in deep models. This leads to an improved understanding of the drawbacks of current methods, and a proposal to improve the quality of uncertainty estimates using probabilistic model ensembles. We give illustrative experiments using MNIST to demonstrate the intuition underlying the different measures of uncertainty, as well as experiments on a real world Kaggle dogs vs cats classification dataset.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("1803.08533")
get_code_for_paper("1803.08533")
have("1803.08533")
Connect an agent — have() is free.