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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Recently more attention has been given to adversarial attacks on neuralnetworks for natural language processing (NLP). A central research topic hasbeen the investigation of search algorithms and search constraints, accompaniedby benchmark algorithms and tasks. We implement an algorithm inspired by zerothorder optimization-based attacks and compare with the benchmark results in theTextAttack framework. Surprisingly, we find that optimization-based methods donot yield any improvement in a constrained setup and slightly benefit fromapproximate gradient information only in unconstrained setups where searchspaces are larger. In contrast, simple heuristics exploiting nearest neighborswithout querying the target function yield substantial success rates inconstrained setups, and nearly full success rate in unconstrained setups, at anorder of magnitude fewer queries. We conclude from these results that currentTextAttack benchmark tasks are too easy and constraints are too strict,preventing meaningful research on black-box adversarial text attacks.

Don't Search for a Search Method -- Simple Heuristics Suffice for Adversarial Text Attacks