# Adversarial Attack on ImageNet (val) (Inception Model Robustness) Source: [Adversarial Attack on ImageNet (val) (Inception Model Robustness)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:02:23.204Z Evaluation of adversarial attack effectiveness on the ImageNet validation split, focusing specifically on the robustness against Inception-family classifiers (Inception-v3, v4, InceptionResNet-v2) and their ensemble variants. 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. | Scope | Recorded value | | --- | --- | | Task | Adversarial Attack | | Dataset (including any stated split) | ImageNet (val) | | Metric | Robust Accuracy (Inc-v3) | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 8 of 8 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Robust Accuracy (Inc-v3)](): Higher is better; unit: SCALAR. - [Robust Accuracy (Inc-v4)](): Higher is better; unit: SCALAR. - [Robust Accuracy (IncRes-v2)](): Higher is better; unit: SCALAR. - [Robust Accuracy (ResNet-101)](): Higher is better; unit: SCALAR. - [Robust Accuracy (Inc-v3 Ensemble 3)](): Higher is better; unit: SCALAR. - [Robust Accuracy (Inc-v3 Ensemble 4)](): Higher is better; unit: SCALAR. - [Robust Accuracy (IncRes-v2 Ensemble)](): Higher is better; unit: SCALAR. ## Recorded results | Method | Robust Accuracy (Inc-v3) | Result date | Report paper | | --- | --- | --- | --- | | NI-CT-FGSM | 1 | Not recorded | 68fa6a921e209131cea28f77 | | VNI-CT-FGSM | 1 | Not recorded | 68fa6a921e209131cea28f77 | | MI-FGSM | 0.999 | Not recorded | 68fa6a921e209131cea28f77 | | VNI-FGSM | 0.999 | Not recorded | 68fa6a921e209131cea28f77 | | NI-FGSM | 0.998 | Not recorded | 68fa6a921e209131cea28f77 | | VMI-FGSM | 0.997 | Not recorded | 68fa6a921e209131cea28f77 | | VMI-CT-FGSM | 0.997 | Not recorded | 68fa6a921e209131cea28f77 | | MI-CT-FGSM | 0.996 | Not recorded | 68fa6a921e209131cea28f77 | ### NI-CT-FGSM Robust Accuracy (Inc-v3): 1. Result ID: 699abf095698665ec44d29c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true","enhanced\_by\_CTM":"true"}. This record does not assert independent reproduction. ### VNI-CT-FGSM Robust Accuracy (Inc-v3): 1. Result ID: 699abf095698665ec44d29c9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true","enhanced\_by\_CTM":"true"}. This record does not assert independent reproduction. ### MI-FGSM Robust Accuracy (Inc-v3): 0.999. Result ID: 699abf095698665ec44d29c2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true"}. This record does not assert independent reproduction. ### VNI-FGSM Robust Accuracy (Inc-v3): 0.999. Result ID: 699abf095698665ec44d29c5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true"}. This record does not assert independent reproduction. ### NI-FGSM Robust Accuracy (Inc-v3): 0.998. Result ID: 699abf095698665ec44d29c4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true"}. This record does not assert independent reproduction. ### VMI-FGSM Robust Accuracy (Inc-v3): 0.997. Result ID: 699abf095698665ec44d29c3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true"}. This record does not assert independent reproduction. ### VMI-CT-FGSM Robust Accuracy (Inc-v3): 0.997. Result ID: 699abf095698665ec44d29c7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true","enhanced\_by\_CTM":"true"}. This record does not assert independent reproduction. ### MI-CT-FGSM Robust Accuracy (Inc-v3): 0.996. Result ID: 699abf095698665ec44d29c6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"attack\_type":"gradient-based iterative","ensemble\_generation":"true","enhanced\_by\_CTM":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.