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

In this work, we attempt to answer a critical question: whether there existssome input sequence that will cause a well-trained discrete-space neuralnetwork sequence-to-sequence (seq2seq) model to generate egregious outputs(aggressive, malicious, attacking, etc.). And if such inputs exist, how to findthem efficiently. We adopt an empirical methodology, in which we first createlists of egregious output sequences, and then design a discrete optimizationalgorithm to find input sequences that will cause the model to generate them.Moreover, the optimization algorithm is enhanced for large vocabulary searchand constrained to search for input sequences that are likely to be input byreal-world users. In our experiments, we apply this approach to dialogueresponse generation models trained on three real-world dialogue data-sets:Ubuntu, Switchboard and OpenSubtitles, testing whether the model can generatemalicious responses. We demonstrate that given the trigger inputs our algorithmfinds, a significant number of malicious sentences are assigned largeprobability by the model, which reveals an undesirable consequence of standardseq2seq training.

Detecting egregious responses in neural sequence-to-sequence models