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

On-device machine learning (ML) is quickly gaining popularity among mobileapps. It allows offline model inference while preserving user privacy. However,ML models, considered as core intellectual properties of model owners, are nowstored on billions of untrusted devices and subject to potential thefts. Leakedmodels can cause both severe financial loss and security consequences. Thispaper presents the first empirical study of ML model protection on mobiledevices. Our study aims to answer three open questions with quantitativeevidence: How widely is model protection used in apps? How robust are existingmodel protection techniques? What impacts can (stolen) models incur? To thatend, we built a simple app analysis pipeline and analyzed 46,753 popular appscollected from the US and Chinese app markets. We identified 1,468 ML appsspanning all popular app categories. We found that, alarmingly, 41% of ML appsdo not protect their models at all, which can be trivially stolen from apppackages. Even for those apps that use model protection or encryption, we wereable to extract the models from 66% of them via unsophisticated dynamicanalysis techniques. The extracted models are mostly commercial products andused for face recognition, liveness detection, ID/bank card recognition, andmalware detection. We quantitatively estimated the potential financial andsecurity impact of a leaked model, which can amount to millions of dollars fordifferent stakeholders. Our study reveals that on-device models are currentlyat high risk of being leaked; attackers are highly motivated to steal suchmodels. Drawn from our large-scale study, we report our insights into thisemerging security problem and discuss the technical challenges, hoping toinspire future research on robust and practical model protection for mobiledevices.

Mind Your Weight(s): A Large-scale Study on Insufficient Machine Learning Model Protection in Mobile Apps