Senior leader at Microsoft explores AI and the phases of transformation for agribusinesses
By Ryan Ridley
Farms.com
Artificial intelligence (AI) is becoming increasingly popular in many industries, and agriculture is no different.
Attendees of the 2020 Farms.com Virtual Precision Agriculture Conference and Ag Technology Showcase learned about AI in agriculture and its implications for agribusinesses with Barney Debnam, Global Agriculture Strategy Lead at Microsoft.
“It’s important to understand what sets AI apart from software is the ability for AI to learn from data and experiences, instead of being explicitly programed in rules,” explains Debnam. “Artificial intelligence is the capability to perceive, learn, reason, and to extend the ability of people and organizations – it’s a set of tools and approaches to augment our own capabilities.”
AI’s capabilities can be grouped into three main categories: system optimization and control, resource and operations optimization, and adaptive experiences.
When it comes to system optimization and control, Debnam uses an example of a combine that has reinforcement learning and adapts in real-time to optimize the harvest experience for a grower.
“Whether that’s changing the platform head or adjusting the internal components for grain loss, [AI] is adapting on the fly,” he adds.
What are AI’s capabilities for resource and operations optimization?
“All the activity that we see in the agribusiness space when we think about nitrogen models, hybrid selection models – these are all about optimizing operations and systems, and certainly some of the leaders in that agribusiness space are pushing the limit when it comes to those items,” says Debnam.
The last category, which Debnam thinks is the most interesting, is adaptive experiences.
How can AI augment people? How can AI help a frontline worker who’s repairing a robotic milker on the farm? What about a planter?
“How can artificial intelligence assist that human being in that repair scenario? There’s a lot of interesting things to drive outcomes – even management decisions as a grower or agribusiness,” he adds.
The artificial intelligence journey for agribusinesses comes in three phases: connected, predictive, and cognitive.
Getting connected deals with defining data and sources, getting plugged in, and seeing what is happening.
“How do I create the data streams to begin to gather information about my business, throughout my business, in a somewhat ubiquitous way, just to create points that may be blind spots in terms of data. This concept of getting connected is fundamental and we’ve seen this evolve in agribusiness at an outstanding pace,” says Debnam.
The predictive phase focuses on historical data and hierarchical data modeling to understand the causes and impacts, as well as AI models and machine learning to predict what will happen next.
“This is all about gaining wisdom on the patterns that might exist,” he explains. “Can I detect when a bearing will fail? Can I detect when that crop is under drought stress, or maybe in an irrigation scenario, that crop has received too much water? Those are all predictive outcomes and are the second step as we see on this journey to cognitive.”
The final phase, cognitive, is when AI takes things to a higher level by teaching, learning, self-reinforcing and gaining insights into how systems should be tuned.
“Highly impactful feedback loops create some cognition about how the system should operate. Perhaps what adjustments or changes should go into the system,” adds Debnam.
He explains that many businesses are at varying stages in this artificial intelligence journey.
“Don’t think about this as all-in for a particular agribusiness – think about it as certain outcomes, certain aspects of the business may move through this continuum at different paces,” he says.
To learn more about artificial intelligence in agribusiness, watch the below video with Debnam.
Photo: microsoft.com