Igor Fernandes, a U of A master's degree student in statistics and analytics who is researching plant breeding through the Center for Agricultural Data Analytics, recently placed first in a contest sponsored by the Sixth International Machine Learning for Cyber-Agricultural Systems workshop. 

The category for the contest was predicting corn yield at a plot level for the 2023 multi-state maize hybrid trials. Researchers at the University of Nebraska and Iowa State University coordinated the competition.

The researchers "generated, described and evaluated over 20,000 plot-level images of over 80 hybrid maize varieties grown across the U.S. corn belt under various management practices collected from (near simultaneous) satellite imagery with ground truth yield measurement."

Field experiments were conducted at six locations, three in Nebraska and three in Iowa. 

"Six satellite images were collected for each location at different dates," said Fernandes. "As satellite images were collected in different locations and dates, I used the Days After Planting (DAP) between the planting date and the date each image was taken as a means to standardize the use of the images."

"The growth stage of the plants might be different across locations because the DAP is different among the locations and across dates of collection," said Fernandes. "For every location, I cherry-picked the "best" satellite image in such a way that, for all locations, the DAPs were very close to the same range (72 to 88 DAP), ensuring that all the plants would be more or less at a similar growth stage. This way, all the images, theoretically, will carry the same information. This part was key."

Fernandes then calculated two vegetation indices commonly used in Precision Agriculture and built a statistical model using these vegetation indices as predictor variables. He also used the hybrid information to try to capture the genetic variability effect.

Twenty-two teams around the world competed in the contest, which ran July 1-Aug. 28. Fernandes receives a $4,000 prize with his first-place finish and will present his findings at the Sixth International Workshop on Machine Learning for Cyber-Agricultural Systems in Lincoln, Nebraska, Oct. 7.

Fernandes is advised by Sam Fernandes, assistant professor of agricultural statistics and quantitative genetics. Fernandes is with the Center for Agricultural Data Analytics and is a researcher and scientist with the Arkansas Agricultural Experiment Station, the research arm of the U of A System Division of Agriculture. His research ties into work with the departments of crop, soil and environmental sciences and horticulture in the Dale Bumpers College of Agricultural, Food and Life Sciences, and he advises graduate students.

The MLCAS workshop focuses on the use of computing technologies to transform traditional plant-based agriculture into an efficient cyber-physical system. The application of advanced Artificial Intelligence and Machine Learning methods can be a transformative extension in agriculture. The workshop brings together academic and industrial researchers and practitioners in the fields of machine learning, data science and engineering, plant sciences and agriculture, in a collaborative effort to identify and discuss technical challenges and recent results related to machine learning-based approaches.

---

About the Dale Bumpers College of Agricultural, Food and Life Sciences: Bumpers College provides life-changing opportunities to position and prepare graduates who will be leaders in the businesses associated with foods, family, the environment, agriculture, sustainability and human quality of life; and who will be first-choice candidates of employers looking for leaders, innovators, policy makers and entrepreneurs. The college is named for Dale Bumpers, former Arkansas governor and longtime U.S. senator who made the state prominent in national and international agriculture. For more information about Bumpers College, visit our website, and follow us on Twitter at @BumpersCollege and Instagram at BumpersCollege.

About the University of Arkansas: As Arkansas' flagship institution, the U of A provides an internationally competitive education in more than 200 academic programs. Founded in 1871, the U of A contributes more than $3 billion to Arkansas' economy through the teaching of new knowledge and skills, entrepreneurship and job development, discovery through research and creative activity while also providing training for professional disciplines. The Carnegie Foundation classifies the U of A among the few U.S. colleges and universities with the highest level of research activity. U.S. News & World Report ranks the U of A among the top public universities in the nation. See how the U of A works to build a better world at Arkansas Research and Economic Development News.