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

Brain-inspired hyperdimensional computing (HDC) is an emerging computationalparadigm that mimics brain cognition and leverages hyperdimensional vectorswith fully distributed holographic representation and (pseudo)randomness.Compared to other machine learning (ML) methods such as deep neural networks(DNNs), HDC offers several advantages including high energy efficiency, lowlatency, and one-shot learning, making it a promising alternative candidate ona wide range of applications. However, the reliability and robustness of HDCmodels have not been explored yet. In this paper, we design, implement, andevaluate HDTest to test HDC model by automatically exposing unexpected orincorrect behaviors under rare inputs. The core idea of HDTest is based onguided differential fuzz testing. Guided by the distance between queryhypervector and reference hypervector in HDC, HDTest continuously mutatesoriginal inputs to generate new inputs that can trigger incorrect behaviors ofHDC model. Compared to traditional ML testing methods, HDTest does not need tomanually label the original input. Using handwritten digit classification as anexample, we show that HDTest can generate thousands of adversarial inputs withnegligible perturbations that can successfully fool HDC models. On average,HDTest can generate around 400 adversarial inputs within one minute running ona commodity computer. Finally, by using the HDTest-generated inputs to retrainHDC models, we can strengthen the robustness of HDC models. To the best of ourknowledge, this paper presents the first effort in systematically testing thisemerging brain-inspired computational model.

HDTest: Differential Fuzz Testing of Brain-Inspired Hyperdimensional Computing