Lucas Waltz ^* Sushma Katari Chaeun Hong Adit Anup Julian Colbert Anirudh Potlapally Taylor Dill Canaan Porter John Engle Christopher Stewart Hari Subramoni Scott Shearer Raghu Machiraju Osler Ortez Laura Lindsey Arnab Nandi Sami Khanal

• The Ohio State University, Columbus, United States

Advancements in machine learning (ML) algorithms that make predictions from data without being explicitly programmed and the increased computational speeds of graphics processing units (GPU) over the last decade have led to remarkable progress in the capabilities of ML. In many fields, including agriculture, this progress has outpaced the availability of sufficiently diverse and high-quality datasets, which now serve as a limiting factor. While many agricultural use cases appear feasible with current compute resources and ML algorithms, the lack of reusable hardware and software components, referred to as cyberinfrastructure (CI) for collecting, transmitting, cleaning, labeling, and training datasets is a major hindrance towards developing solutions to address agricultural use cases. This study aims to share the learnings associated with collecting, processing, and training ML models for three agricultural use cases derived from a 1-terabyte (TB) multimodal dataset from three agricultural research locations across Ohio during the 2023 growing season. The dataset includes Unmanned Aerial System (UAS) imagery (RGB and multispectral), and soil and weather sensors for the state's two most widely cultivated crops: corn and soybean. Based on these learnings, this study presents a vision for agriculture-focused CI that can amplify the efforts of agricultural researchers and lower the barriers to developing ML applications in agriculture.