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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In security-sensitive applications, the success of machine learning dependson a thorough vetting of their resistance to adversarial data. In onepertinent, well-motivated attack scenario, an adversary may attempt to evade adeployed system at test time by carefully manipulating attack samples. In thiswork, we present a simple but effective gradient-based approach that can beexploited to systematically assess the security of several, widely-usedclassification algorithms against evasion attacks. Following a recentlyproposed framework for security evaluation, we simulate attack scenarios thatexhibit different risk levels for the classifier by increasing the attacker'sknowledge of the system and her ability to manipulate attack samples. Thisgives the classifier designer a better picture of the classifier performanceunder evasion attacks, and allows him to perform a more informed modelselection (or parameter setting). We evaluate our approach on the relevantsecurity task of malware detection in PDF files, and show that such systems canbe easily evaded. We also sketch some countermeasures suggested by ouranalysis.

Evasion Attacks against Machine Learning at Test Time