# Adversarially robust transfer learning Source: [Adversarially robust transfer learning]() Collection index: [llms.txt]() Data updated: 2025-10-25T22:06:19.934Z Authors: Ali Shafahi, Parsa Saadatpanah, Chen Zhu, Amin Ghiasi, Christoph Studer, David Jacobs, Tom Goldstein Published: 2019-05-20T17:57:57.000Z ## Abstract Transfer learning, in which a network is trained on one task and re-purposed on another, is often used to produce neural network classifiers when data is scarce or full-scale training is too costly. When the goal is to produce a model that is not only accurate but also adversarially robust, data scarcity and computational limitations become even more cumbersome. We consider robust transfer learning, in which we transfer not only performance but also robustness from a source model to a target domain. We start by observing that robust networks contain robust feature extractors. By training classifiers on top of these feature extractors, we produce new models that inherit the robustness of their parent networks. We then consider the case of fine tuning a network by re-training end-to-end in the target domain. When using lifelong learning strategies, this process preserves the robustness of the source network while achieving high accuracy. By using such strategies, it is possible to produce accurate and robust models with little data, and without the cost of adversarial training. Additionally, we can improve the generalization of adversarially trained models, while maintaining their robustness. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Classification on CIFAR100 (Adversarial Robustness - Clean & PGD)](): Image Classification; CIFAR100; Clean Accuracy: 74.36. - [Adversarial Robustness on CUB200 (Image Classification)](): Image Classification; CUB200; Accuracy (Clean): 70.83. - [Adversarial Robustness Evaluation on Stanford Dogs (Image Classification)](): Image Classification; Stanford Dogs; Clean Accuracy: 91.55. - [Image Classification on CIFAR10 (Clean and PGD Robustness)](): Image Classification; CIFAR10; Clean Accuracy: 93.02. - [Image Classification on Caltech256 (Robustness Metrics)](): Image Classification; Caltech256; Accuracy (Clean): 81.79. ## Original sources and implementation [Original paper]() [PDF]()