Pretraining Probabilistic Models for Scalable Precision Agriculture
Keywords: Transfer Learning, Precision Agriculture
Abstract: Genomic predictions can help breeders select desirable plant traits, improving crop yields and resilience. However, data collection for developing these prediction models is expensive. Using low-cost auxiliary data that exhibit correlation with desired traits can reduce costs and ultimately improve prediction accuracy. Although such data are abundant, identifying meaningful auxiliary candidates is time-consuming. To this end, we propose a transfer learning mechanism on Gaussian processes to search for potential good candidates via Bayesian optimization. Our results demonstrate promising transferability, paving a new way for efficient searching with a parsimonious sample size.
Primary Subject Area: Domain specific data issues
Paper Type: Research paper: up to 8 pages
DMLR For Good Track: Participate in DMLR for Good Track
Participation Mode: In-person
Confirmation: I have read and agree with the workshop's policy on behalf of myself and my co-authors.
Submission Number: 100
Loading