Explainable Zero Shot Honey Adulteration Detection using β- Variational Autoencoders and Hyperspectral Imaging

Honey fraud is a rising global issue, with recent investigations by the authenticity networks revealing that over 90% of honey samples from major retailers (such as from the UK) are laced with cheap fillers. These fillers, essentially syrups derived from rice, beetroot and corn are fabricated to mimic the physical properties of natural nectar while lacking its nutritional and enzymatic complexity. This high-end adulteration is driven by a stark economic disparity with authentic pure honey valued at $3,000 per tonne compared to just $500 per tonne for synthetic syrups. Be- yond consumer deception, this cheap adulteration not only threatens the economic viability of beekeeping, but also poses direct threats to global biodiversity as well as crop pollination processes.