Development of a milk (mid-infrared) MIR Spectral deep learning neural network as a classifier model for GHG emission profiles in dairy cattle
Prof Craig Michie, University of Strathclyde; Prof Chris Creevey, Queen’s University Belfast; Dr Mazdak Salavati, SRUC
Research Aims
The project focuses on using artificial intelligence (AI) in the dairy sector to identify novel traits with strongest degrees of biological connection to Green House Gas (GHG) emissions from animals. Most dairy farms in the UK and around the world sample the milk on a regular basis for fat, protein, and spectral measurements. A recent technology which has been taken up globally is milk mid-infrared spectroscopy or MIR.
The project aims to develop a milk mid-infrared (MIR) spectral deep learning neural network as a classifier model for GHG emission profiles in dairy cattle.
The goal is to create a standardization pipeline for MIR spectral datasets generated across different herds for cross-validation, optimize and deploy an artificial neural network (ANN) for predicting highly associated datapoints with health and production traits in two dairy cattle populations. This pipeline will be fine-tuned for reproducibility within GPU-base environments.
This multi-faceted approach will harness the power of digital transformation and field applications of AI within the dairy sector to tackle the urgent issue of reducing the carbon footprint of the dairy supply chain.
The successful student will learn both data analytics and software development skills as part of this project with a focus on real world application of AI in the dairy sector.