Applying machine learning models to improve dairy farm management

By Meghan Chua


As new technologies have opened doors for dairy farms to harness more data from their herds than ever before, farmers around the state have embraced these innovations.

That leaves farmers with vast amounts of data – on cows, herds, farms, the market, crops, and soils – but, as of yet, no way to integrate the entirety of that data into farm management.

Currently, a team of researchers at UW–Madison is working to create a virtual dairy farm brain that can apply machine learning and artificial intelligence to dairy farm management to help farmers make better decisions. The project, funded by the UW2020 initiative, is led by Victor Cabrera, associate professor of dairy science and UW-Extension dairy systems management specialist, with collaborators in the departments of Agricultural and Applied Economics, Computer Sciences, and UW-Extension.

Dairy cows are pictured on UW alum Mitch Bruenig's farm near Roxbury

Computer Sciences graduate student Anuja Golechha is among the project’s research assistants. With an interest in machine learning and artificial intelligence, the project piqued Anuja’s interest when her advisor, Jignesh Patel, recommended it.

The project is particularly relevant to Wisconsin, where the dairying industry contributes half of the agricultural economy, supports 80,000 jobs, and has an impact of $43 billion a year.

Anuja works on one piece of the overall project. Her focus is mastitis, a disease that hinders cows’ milk production. Using factors such as the cow’s medial history, its lactation period, and whether it has had mastitis before, Anuja is building a model to predict the likelihood of mastitis occurrence in the future.

Currently, she’s building a temporal model focused on milk production levels, which drop prior to a mastitis diagnosis.

Though Anuja admits that she is not a dairy scientist, she’s certainly been learning about these topic as part of the project, relying on other researchers and postdocs in the Dairy Science program to make sure the model is informed by the right factors.

Anuja said that working on the project is a great chance for her to work with real-world data.

“For a machine learning student, it’s important to work with real-world data and find out what the problems are,” she said.

The UW2020 initiative supports innovative and groundbreaking research at the University of Wisconsin–Madison with the potential to transform a field of study. UW2020 grants are supported by the Wisconsin Alumni Research Foundation (WARF) with combined funding from the Graduate School and other sources.