English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Autonomous systems are highly vulnerable to a variety of adversarial attackson Deep Neural Networks (DNNs). Training-free model-agnostic defenses haverecently gained popularity due to their speed, ease of deployment, and abilityto work across many DNNs. To this end, a new technique has emerged formitigating attacks on image classification DNNs, namely, preprocessingadversarial images using super resolution -- upscaling low-quality inputs intohigh-resolution images. This defense requires running both image classifiersand super resolution models on constrained autonomous systems. However, superresolution incurs a heavy computational cost. Therefore, in this paper, weinvestigate the following question: Does the robustness of image classifierssuffer if we use tiny super resolution models? To answer this, we first reviewa recent work called Super-Efficient Super Resolution (SESR) that achievessimilar or better image quality than prior art while requiring 2x to 330x fewerMultiply-Accumulate (MAC) operations. We demonstrate that despite being ordersof magnitude smaller than existing models, SESR achieves the same level ofrobustness as significantly larger networks. Finally, we estimate end-to-endperformance of super resolution-based defenses on a commercial Arm Ethos-U55micro-NPU. Our findings show that SESR achieves nearly 3x higher FPS than abaseline while achieving similar robustness.

Super-Efficient Super Resolution for Fast Adversarial Defense at the Edge