Deep multi-task learning for stem water potential prediction: actionable guidelines

Abstract

Predicting stem water potential (ψ _stem) accurately into the future remains a significant challenge in precision agriculture, vital for optimal irrigation and sustainable crop health management. This paper addresses this challenge using advanced machine learning techniques, specifically focusing on the recently developed MultiMix method — a multi-task learning framework for limited data environments. This paper investigates MultiMix's ability to forecast ψ _stem – given limited ψ _stem-measurements – for different crops. We conduct a thorough evaluation of its performance under various experimental set-ups and offer practical guidelines for implementation in agriculture. The results demonstrate that MultiMix effectively handles the complexities inherent in forecasting ψ _stem for different crops and reveals interesting insights for practitioners.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata

Altmetrics