Higher-Order Certification For Randomized Smoothing
Authors: Jeet Mohapatra, Ching-Yun Ko, Tsui-Wei Weng, Pin-Yu Chen, Sijia Liu, Luca Daniel
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | On CIFAR10 and Imagenet datasets, the new regions certified by our approach achieve significant improvements... We empirically study the performance of the new certification schemes on standard image classification datasets, CIFAR10 and Imagenet. |
| Researcher Affiliation | Collaboration | Jeet Mohapatra1 Ching-Yun Ko1 Tsui-Wei Weng1,2 Pin-Yu Chen2 Sijia Liu2 Luca Daniel1 1 MIT 2 MIT-IBM Watson AI Lab, IBM Research |
| Pseudocode | No | Table 1: Estimators to calculate the different norm value of g(x). (* newly designed estimators) |
| Open Source Code | No | The paper does not provide any specific links or explicit statements regarding the availability of its source code. |
| Open Datasets | Yes | On CIFAR10 and Imagenet datasets, the new regions certified by our approach achieve significant improvements... We empirically study the performance of the new certification schemes on standard image classification datasets, CIFAR10 and Imagenet. |
| Dataset Splits | No | For both our proposed certificate and the baseline certificate [5], we use a failure probability of α = 0.001 and N = 200, 000 samples for CIFAR10 and N = 1, 250, 000 samples for Imagenet. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers. |
| Experiment Setup | Yes | For both our proposed certificate and the baseline certificate [5], we use a failure probability of α = 0.001 and N = 200, 000 samples for CIFAR10 and N = 1, 250, 000 samples for Imagenet. ...In our plots, we present, for each threat model the upper envelopes of certified accuracies attained over the range of considered σ {0.12, 0.25, 0.50, 1.00}. |