1
HiveGuard Research Initiative
ProposalGrant
Closes May 22nd, 2026$0raised
$500minimum funding
$48,000funding goal
Offer to donate
29 daysleft to contribute
You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
Project Summary
HiveGuard is developing a non-invasive, AI-driven acoustic sensor to detect Varroa destructor mites in honeybees without killing a single bee. Currently, the industry standard ("alcohol wash") requires killing ~300 bees per sample, creating a moral barrier to early detection and forcing reliance on toxic chemical miticides. HiveGuard uses machine learning to isolate mite-specific acoustic signatures, enabling early, humane interventions (heat/vibration) that prevent colony collapse.
Goals & Methodology
Primary Goal: Validate a prototype sensor that predicts mite loads with >90% accuracy against gold-standard diagnostics.
Methodology: Deploy high-fidelity microphones in hives at UC Davis (3-4 colonies committed by Dr. Elina Niño) and the National Bee Diagnostic Centre (Canada). Collect 500+ hours of audio data, train an Attention-Based Neural Network, and correlate acoustic patterns with mite counts.
Outcome: A validated, open-source dataset and "Welfare Indicator Matrix" that enables non-chemical pest control in commercial apiaries.
Use of Funds
We are seeking $48,000 from the Survival and Flourishing Fund (SFF) to fund a 6-month pilot (April–September 2026):
$10,000: Technical Contractor (Signal Processing/AI).
$16,000: PI Stipend (Ray Hsu, Ph.D.).
$7,000: Hardware Prototyping (Sensors, enclosures).
$8,000: Field Stipends (UC Davis/NBDC partners).
$6,000: Compute & Data Costs.
$5,000: Dissemination (Publication & Open Access).
Team & Track Record
PI: Ray Hsu, Ph.D. (AI/ML Product Leader, Sentient Futures Fellow, former UBC faculty). Expertise in AI ethics, systems design, and project management.
Scientific Validators: Dr. Elina Niño (UC Davis) and Dr. Steve Pernal (National Bee Diagnostic Centre, Canada).
Advisors: Elisabeth Ormandy (Animal Charity Evaluators), Max Taylor (SFF Recommender).
Track Record: Successfully led the development of a VR-based high school dissection alternative (Animals in Science/CSHS), replacing animal use in education.
Failure Modes & Mitigation
Risk: Acoustic signals are too noisy to isolate mites.
Mitigation: We are leveraging a 2023 Scientific Reports (Nature) study that proved mite gait signatures exist. Our AI model is designed specifically for temporal filtering to overcome noise.
Risk: Pilot data is insufficient for validation.
Mitigation: We have committed access to 3-4 colonies at UC Davis and are engaging the NBDC for federal validation, ensuring high-quality ground-truth data.
Outcome if failed: We will publish the negative results and open-source the dataset to prevent others from repeating the experiment, contributing to the field's knowledge base.
Funding History
Last 12 Months: $0 raised. This is a new R&D initiative.
Current Status: Applying for the SFF Main Round (Deadline: April 22, 2026). Pending fiscal sponsorship confirmation.