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

It has been proved that deep neural networks are facing a new threat calledbackdoor attacks, where the adversary can inject backdoors into the neuralnetwork model through poisoning the training dataset. When the input containingsome special pattern called the backdoor trigger, the model with backdoor willcarry out malicious task such as misclassification specified by adversaries. Intext classification systems, backdoors inserted in the models can cause spam ormalicious speech to escape detection. Previous work mainly focused on thedefense of backdoor attacks in computer vision, little attention has been paidto defense method for RNN backdoor attacks regarding text classification. Inthis paper, through analyzing the changes in inner LSTM neurons, we proposed adefense method called Backdoor Keyword Identification (BKI) to mitigatebackdoor attacks which the adversary performs against LSTM-based textclassification by data poisoning. This method can identify and excludepoisoning samples crafted to insert backdoor into the model from training datawithout a verified and trusted dataset. We evaluate our method on fourdifferent text classification datset: IMDB, DBpedia ontology, 20 newsgroups andReuters-21578 dataset. It all achieves good performance regardless of thetrigger sentences.

Mitigating backdoor attacks in LSTM-based Text Classification Systems by Backdoor Keyword Identification