Back

 Industry News Details

 
AI Is Learning To Understand How Vegetables Taste Posted on : Jul 21 - 2021

With the global demand for food escalating, vertical farms are becoming a critical component of agriculture's future. They use robotics, machine learning and artificial intelligence (AI) to automate farming and perfect the growing of greens and vegetables. With steady growth, the vertical farming market was had an estimated value of $4.4 billion in 2019 and is expected to reach $15.7 billion by 2025.

Fifth Season, a vertical farm in Pittsburgh with $35 M in funding, uses super-stack software and robotics to run their fully automated farming systems. And, by combining big data and AI, they have created the optimal grow recipe that determines the best flavor for the plants they grow.

"The role of AI in determining flavor is to leverage big data and AI to ensure you achieve the target flavor — sweetness, spiciness, bitterness, total degree of flavor and texture," said Austin Webb, CEO of Fifth Season.

"Our plant's individualized grow recipe is the unique mix of the different LED lights," said Webb. "The plants go through the grow room with a QR code that communicates that plant's route and tells the automated system where each plant needs to be throughout the process."

Webb says their super stack system, which serves as 'the brain' of the vertical farm maps, maps out each plant's route through the grow room based on its grow recipe and then moves the plants where they need to go.

"We use AI and data to find improvements in all aspects of crop quality, even beyond what humans think they know about flavor profiles. We call this proactive, deterministic growing compared to traditional farming, including greenhouse growing, where you have to be reactive based on weather and sunlight conditions," said Webb. "We then leverage human/chef feedback on what tastes the best and what texture is best and [..] combine that qualitative data with the 26K quantitative data points for every tray of greens per lifecycle."

"From there, we tweak our grow recipes to build the best flavor. For some vegetables, like tomatoes, experts have leveraged Brix scores, but [..] we measure flavor quality based on these factors: sweetness, spiciness, bitterness, the total degree of flavor, texture and color," said Webb. "Humans don't need to guess what iron content or Brix score is best; the brain in our farms can do that. Humans tell the grain what tastes best, and the brain will compute and tweak the grow recipes from there."

Darryn Keiller, CEO and founder of WayBeyond, says that to impact flavor, you either have to change the genetics of the crop or alter the existing biochemical profile. View More