# Image Classification on ImageNet-1K (val) (Certified Robustness Evaluation) Source: [Image Classification on ImageNet-1K (val) (Certified Robustness Evaluation)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:41:58.650Z Evaluation of Image Classification performance on the validation split of ImageNet-1K, focusing specifically on certified robustness metrics, including Empirical Accuracy and Certified Accuracy across various perturbation levels (e.g., epsilon 0.0 to 2.0). 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 | Image Classification | | Dataset (including any stated split) | ImageNet IN-1K (val) | | Metric | Empirical Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 7 of 7 | | 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 - [Empirical Accuracy](): Higher is better; unit: PERCENT. - [Certified Accuracy (eps=0.0)](): Higher is better; unit: PERCENT. - [Certified Accuracy (eps=0.5)](): Higher is better; unit: PERCENT. - [Certified Accuracy (eps=1.0)](): Higher is better; unit: PERCENT. - [Certified Accuracy (eps=1.5)](): Higher is better; unit: PERCENT. - [Certified Accuracy (eps=2.0)](): Higher is better; unit: PERCENT. - [ACR](): Higher is better; unit: SCALAR. ## Recorded results | Method | Empirical Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Diffusion Calibration | 83.8 | Not recorded | 68fa710a1e209131cea4ed24 | | Cascaded Randomized Smoothing | 83.5 | Not recorded | 68fa710a1e209131cea4ed24 | | Diffusion Calibration | 83.2 | Not recorded | 68fa710a1e209131cea4ed24 | | Cascaded Randomized Smoothing | 82.6 | Not recorded | 68fa710a1e209131cea4ed24 | | Carlini et al. | 80.9 | Not recorded | 68fa710a1e209131cea4ed24 | | Carlini et al. | 79.5 | Not recorded | 68fa710a1e209131cea4ed24 | | Carlini et al. | 77.2 | Not recorded | 68fa710a1e209131cea4ed24 | ### Diffusion Calibration Empirical Accuracy: 83.8. Result ID: 699190963a6c3148490902c2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"sigma":"(0.25-0.50)","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. ### Cascaded Randomized Smoothing Empirical Accuracy: 83.5. Result ID: 699190963a6c3148490902c1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"sigma":"(0.25-0.50)","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. ### Diffusion Calibration Empirical Accuracy: 83.2. Result ID: 699190963a6c3148490902c5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"sigma":"(0.25-1.00)","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. ### Cascaded Randomized Smoothing Empirical Accuracy: 82.6. Result ID: 699190963a6c3148490902c4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"sigma":"(0.25-1.00)","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. ### Carlini et al. Empirical Accuracy: 80.9. Result ID: 699190963a6c3148490902bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"sigma":"0.25","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. ### Carlini et al. Empirical Accuracy: 79.5. Result ID: 699190963a6c3148490902c0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"sigma":"0.50","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. ### Carlini et al. Empirical Accuracy: 77.2. Result ID: 699190963a6c3148490902c3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"sigma":"1.00","Backbone":"ViT-B/16@224"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.