a type of pneumonia - is the leading cause of death for dairy calves after they become accustomed to food other than their mothers' milk, resulting in economic losses at over $1 billion annually for the U.S. cattle industry. To detect BRD in dairy calves before they show obvious symptoms and reduce those costly losses, a team of researchers at Penn State, the University of Kentucky and the University of Delaware, funded by a new three-year, $1 million grant from the U.S. National Science Foundation, intend to create a system that uses modern sensing technologies and advanced artificial intelligence (AI).
"We know that early detection can save the lives of calves, reduce antibiotic use and improve farmers' profitability," said Melissa Cantor, assistant professor of precision dairy science and lead collaborator at Penn State, in the College of Agricultural Sciences. "So, we will build an explainable, affordable and widely applicable AI system called CalfHealth to detect calf pneumonia early using wearable sensors and robotic smart feeders, among other tools. We'll also develop an understanding of how to get farmers to adopt and trust such a system that combines computer science, animal science, economics and behavioral science."
The other co-principal investigators on the project are Simone Silvestri, professor in the Department of Computer Science at the University of Kentucky, and Michelle Segovia, associate professor of food and agribusiness marketing at the University of Delaware. The work is a continuation of research Cantor and Silvestri published two years ago.
The CalfHealth system will incorporate several innovations, according to Cantor. It will feature multimodal detection, meaning it will monitor several data types, combining behavior data from low-cost sensors calves will wear, such as accelerometers that track the young animals' steps and lying down resting times. The plan is that the system will observe feeding behavior from precision robotic feeders, and it will detect breathing patterns using a non-invasive and inexpensive Wi-Fi-based sensing system.
The AI will use a deep-learning approach called attention mechanisms to determine which behaviors or breathing changes matter most for detecting illness at any given time. The system will incorporate other types of AI learning so that it can immediately work across many farms despite differences in size, layout and management styles, Cantor noted, a feature she said is critical for real-world adoption.
Cantor noted that CalfHealth could be of use beyond dairy farming - for example, for beef cattle, other calf diseases such as diarrhea and detecting early signs of emerging epidemics such as avian influenza - but farmers need to be able to use it first.
"To enhance explainability, the system will include an interactive, farmer-facing chatbot powered by a selection of language models," she said, explaining that users need to understand the technology and how it works before they can trust that it will help manage their livestock. "The chatbot will explain why the system flagged a calf as at risk and allow farmers to ask questions, including a 'what-if' analysis, such as, 'What if a calf's feeding drops but its breathing stays normal?'"
The researchers plan to study how to build trust between farmers and AI tools through behavioral science experiments. They also will test CalfHealth on multiple farms, measuring the impact of early detection on calf health and farm profitability. Finally, they will evaluate what encourages farmers to adopt such technologies or keeps them from doing so. The research project will include workshops for farmers, veterinarians and industry stakeholders, as well as presentations to local communities and demonstrations of the technology.
At Penn State, researchers are solving real problems that impact the health, safety and quality of life of people across the commonwealth, the nation and around the world.
For decades, federal support for research has fueled innovation that makes our country safer, our industries more competitive and our economy stronger. Recent federal funding cuts threaten this progress.
Learn more about the implications of federal funding cuts to our future at Research or Regress.
