Adversarially Robust Conformal Prediction

Authors: Asaf Gendler, Tsui-Wei Weng, Luca Daniel, Yaniv Romano

ICLR 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the performance of our proposed methods on three benchmark image classification data sets: CIFAR10, CIFAR100 (Krizhevsky, 2009), and Image Net ILSVRC2012 (Deng et al., 2009), which are described in Supplementary Section S2.
Researcher Affiliation Academia 1Department of Electrical and Computer Engineering, Technion Israel Institute of Technology 2Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
Pseudocode Yes Algorithm 1 RSCP: Randomly Smoothed Conformal Prediction
Open Source Code Yes A Python package that implements our methods and code to reproduce our experiments are available at https://github.com/Asafgendler/RSCP.
Open Datasets Yes We evaluate the performance of our proposed methods on three benchmark image classification data sets: CIFAR10, CIFAR100 (Krizhevsky, 2009), and Image Net ILSVRC2012 (Deng et al., 2009), which are described in Supplementary Section S2.
Dataset Splits Yes Then, we split the remaining data into two equally sized disjoint subsets, one is used for calibration and the second for testing.
Hardware Specification Yes We train the models using Pytorch, on a single Nvidia GEFORCE GTX 1080 Ti GPU.
Software Dependencies No We train the models using Pytorch, on a single Nvidia GEFORCE GTX 1080 Ti GPU. This mentions Pytorch but does not specify a version number or other software dependencies with versions.
Experiment Setup Yes For all data sets, we choose Res Net (He et al., 2016) to be the base architecture of the deep net classifier and fit two different models for each data set: one on clean training points and the second on points that are augmented with Gaussian noise of standard deviation σ that is equal to the smoothing parameter from (10).