Preliminary Agenda

Monday, April 15th

Multi-state S1090 Meeting
10:30 a.m. – 11:30 a.m.

Registration (Foyer)
12:00 p.m. – 6:00 p.m.

Prairie View A&M University Tour
12:00 p.m. – 5:00 p.m.

Pre-conference Workshop 1 (Ballroom)
12:00 p.m. – 1:00 p.m.
Josh Peeples

Plant phenotyping has a profound impact on real-world problems such as food security and energy demands (i.e., biofuel). Recent advances in artificial intelligence (AI) have been successfully integrated to automate plant phenotyping, resulting in reductions of human error and improved efficiency. Texas A&M AgriLife Research’s new state-of-the-art facility captures high-quality, multi-spectral images of various crops. To achieve the goal of automated plant phenotyping, a centralized data hub is being developed to achieve the goal of automated plant phenotyping. In this workshop, we will briefly explore each module of the data pipeline and provide interactive examples of working with the data acquired in the facility.

Pre-conference Workshop 2: XGBOOST Your Digital Ag Research (Ballroom)
1:00 p.m. - 3:00 p.m.
Thanos Gentimis, Dina Dinh, Leticia Santos

In this workshop, we will explore the new and promising Machine Learning paradigm of XGBOOST on an information-rich agriculture-based dataset. Our goal will be to predict various agronomic indices, including yield and vigor through a Python-based code that compares multiple models, with an emphasis on the appropriate use and optimization of those models. This workshop will be using Google Collab as its primary delivery platform, but there will be optional videos available to the audience that will enable them to port the workshop to their own Python-based platforms. No advanced coding experience nor knowledge of the specific models used is required, and all data will be provided, but you will need to bring your own laptops. We encourage the participants to ask questions, and a Github repository will be set up for all FAQs, codes, and recordings.

Pre-conference Workshop 3: Fundamentals of Image Classification with a Simple Neural Network Using the Keras API (Ballroom)
3:00 p.m. - 4:00 p.m.
Dr. Yalong Pi

Participants will focus on fundamental aspects of image classification using the Keras API with a handwritten digit dataset. They will learn key steps including image (matrix) structures, dataset loading and preparation, neural network construction for classification, and model training. Through hands-on exercises, participants will understand the fundamental components of machine learning especially how the neural network trains, laying the foundation for the next workshop which applies convolutional neural networks (CNN) for a practical problem: cotton water stress detection. Finally, they will see how to deploy the trained model and interpret the output of image classification problems.

Pre-conference Workshop 4: End-to-End Workflows for CNN Models in Agriculture: A Case Study on UAV Image-Based Water Stress Classification in Cotton (Ballroom)
4:00 p.m. - 5:00 p.m.
Dr. Haoyu Niu

This workshop introduces participants to end-to-end workflow for data analysis and convolution neural network (CNN) modeling, through hands-on computational experience in application to cotton water stress classification. The case study uses cotton RGB images collected by a UAV system at Lubbock, Texas, comprising 6,832 images for each collection date. This workshop leads participants through the stages of the data workflow: data preprocessing, exploratory data analysis, building CNN models from scratch, hyperparameter tuning with KerasTuner, CNN model training and evaluation, and data visualization. Participants will emerge from the workshop equipped with comprehensive skills to leverage fundamental data science, machine learning, and deep learning knowledge for their own applications. The earlier workshop “Fundamentals of Image Classification with a Simple Neural Network using the Keras API.” is preparatory for this workshop.

Welcome (Ballroom)
5:00 p.m. – 5:30 p.m.

Keynote 1 (Ballroom)
5:00 p.m. - 6:30 p.m.
Emre Kiciman & Ranveer Chandra — Microsoft

Dinner & Networking (TBD)
6:30 p.m. – 8:00 p.m.

Tuesday, April 16th

Registration & Breakfast (Foyer)
7:30 a.m. – 8:30 a.m.

Keynote Speaker: AI Foundation Models for Agriculture (Ballroom)
8:30 a.m. – 9:30 a.m.
Hendrick Hamann

In recent years, the landscape of artificial intelligence (AI) has been reshaped by the rapid emergence of Foundation Models (FMs). These versatile models have garnered widespread attention for their remarkable ability to transcend the boundaries of traditional, bespoke AI solutions and to generalize to a large set of downstream tasks. Consequently, a potentially game-changing advantage of such FMs lies in the fact that they scale across different application domains while being very performant both in accuracy and compute efficiency. In this presentation, we will introduce this technology and show examples of the enormous opportunities for FMs to accelerate the development of solutions relevant to agriculture, climate adaptation (e.g., climate risks and impacts), and mitigation (e.g., nature-based carbon sequestration).

Coffee Break (Foyer)
9:30 a.m. – 9:50 a.m.

Five Concurrent Breakout Sessions

Meeting rooms: Ballroom, Corps, Ross, Reveille, Traditions

9:50 a.m. – 10:10 a.m.

• Session 1A: Sensors & Robotics (Ballroom) — Image Repository for Cotton, Corn, and Sorghum Production Using the Farming amiga robot.
  Oscar Fernandez
• Session 1B: Pests & Diseases (Corps) — Developing an AI-based Plant Weed Detector for Smart Agriculture.
  Ahmed Ahmed
• Session 1C: Social & Economic Implications (Ross) — Ag Econ Smart (AES): A Suite of Agricultural Economics Online Tools.
  Nazia Arbab
• Session 1D: Water & Soils (Reveille) — A regression-based approach to estimate soil water content in cover crop-based cotton production systems from UAS-based images.
  Srinivasulu Ale
• Session 1E: Management – Yield (Traditions) — Image and cyber-physical systems-based high throughput phenotyping aimed at improving crop productivity
  Krishna Jagadish

10:10 a.m. – 10:30 a.m. 

• Session 1A: Sensors & Robotics (Ballroom) — Hierarchical Reinforcement Learning for Autonomous Harvesting Robots in Dynamic Environments
  Prasad Nethala
• Session 1B: Pests & Diseases (Corps) — From Pests to Prevention: Leveraging mmwave Radar Sensor and Machine Learning for Proactive Pest Management
  Leslie Barreto Gonzalez
• Session 1C: Social & Economic Implications (Ross) — Behavioral and Technological Drivers of Adoption of Digital Twins in Agriculture
  Michelle Segovia
• Session 1D: Water & Soils (Reveille) — Unfolding the Complexity of Soil: Harnessing AI and Remote Sensing for Cutting-Edge Detection, Monitoring, and PhotoAOP Remediation
  Mishaal Ashkanani
• Session 1E: Management – Yield (Traditions) — Improving Yield Prediction Accuracy with Machine Learning and Field Boundary Effects
  Thanos Gentimis

10:30 a.m. – 10:50 a.m.

• Session 1A: Sensors & Robotics (Ballroom) — Development of a machine vision and spraying system of a robotic precision smart sprayer for specialty crops.
  Vinay Vijayakumar
• Session 1B: Pests & Diseases (Corps) — Weed Detection in Early-Stage Sugarcane Fields Using Machine Learning and Low-Resolution RGB Imagery.
  Lalita Panduangnat
• Session 1C: Social & Economic Implications (Ross) — Using Eye-Tracking to Understand Cattle Producers Take-Up Decision.
  Christopher Boyer
• Session 1D: Water & Soils (Reveille) — Projection of Hydrologic Intensity Duration Frequency Parameters and Their Uncertainties Based on Climate Projections for the 21st Century in the State of Texas.
  Fouad Jabar
• Session 1E: Management – Yield (Traditions) — Machine Learning Based Prediction of Wheat Yield using Farmer's Surveys data.
  Siraj Mohammed

10:50 a.m. – 11:10 a.m.

• Session 1A: Sensors & Robotics (Ballroom) — Performance Benchmarking of Deep Object Detectors to Detect Green Asparagus Towards Robotic Harvesting.
  Jiajun Xu
• Session 1B: Pests & Diseases (Corps) — Drone-Based Intelligent for the Management of Pests in Crops: A Comprehensive Study.
  Shreya Singh
• Session 1C: Social & Economic Implications (Ross) — Blockchain-Enhanced Data Management in AI-Driven Agriculture: A Pathway to Efficiency and Transparency.
  Younghoo Cho
• Session 1D: Water & Soils (Reveille) — 3-Year Agricultural Field Study for GNSS-R Based Soil Moisture Mapping.
  Volkan Senyurek
• Session 1E: Management – Yield (Traditions) — Soybean Flower Abortion Phenotyping via Image Analysis and Machine Learning Techniques
  Juliana Espindola

11:10 a.m. – 11:30 a.m.

• Session 1A: Sensors & Robotics (Ballroom) — Streamlining UAV Data Analysis in Agriculture through a Comprehensive Canopy Feature Database.
  Jose Luis Landivar
• Session 1B: Pests & Diseases (Corps) — Detecting citrus pests from sticky traps using deep learning.
  Congliang Zhou
• Session 1C: Social & Economic Implications (Ross) — Evaluating the Performance Extension Bot: An Agriculturally Focused Small Language Chatbot for Land Grant Extension Services.
  Jeffrey Vitale
• Session 1D: Water & Soils (Reveille) — Enhancing Agricultural Water Management: A Desktop Application Integrating UAV Imagery and Ground Sensing for Precision Irrigation.
  Boaz Tulu
• Session 1E: Management – Yield — Modelling root zone soil moisture to be a searchable chart for farming practices in the Southeast US.
  Ziwen Yu

Lunch (Ballroom)
11:30 p.m. – 12:30 p.m.

Panel 1 (Ballroom)
12:30 p.m. – 1:45 p.m.

Challenges and Opportunities in the adaptation of digital agriculture technology by producers.

• Dr. Kater Hake (moderator)
• Mr. Bob Walker (Producer)
• Ms. Lacey Vardeman (Producer)
• Mr. Tracey Senter (Producer)
• Ms. Kelly Whatley  (Producer)

TAMU Phenotyping Greenhouse Tour (will run concurrently with the panel)
12:30 p.m. - 1:30 p.m.
Bus to Phenotyping Greenhouse, meet in pre-conference space (5-minute ride).

Coffee Break (Foyer)
1:45 p.m. – 2:00 p.m. 

Five Concurrent Breakout Sessions

Meeting rooms: Ballroom, Corps, Ross, Reveille, Traditions

2:00 p.m. – 2:20 p.m.

• Session 2A: Breeding & Phenotyping (Ballroom) — Transforming Advanced Machine Learning Algorithms into User-Friendly Tools for Breeding.
  Heather Manching
• Session 2B: Livestock (Corps) — Designing a Hybrid Intelligent Decision Support System for Sustainable Beef Production: The Case of Bovine Respiratory Disease and Enteric Emissions Mitigation.
  K. Kaniyamattam
• Session 2C: Education & Outreach (Ross) — Cross-Disciplinary Mentoring for Undergraduates: Data Analytics in Agriculture.
  Yuxia (Lucy) Huang
• Session 2D: Horticulture – Field (Reveille) — A data-driven approach for precision picker activity recognition during manual fruit harvesting.
  Uddhav Bhattarai
• Session 2E: Management Cotton (Traditions) — Artificial Intelligence and Satellite-Based Remote Sensing Model to Predict Cotton (Gossypium spp.) YIELD.
  Dulis Duron

2:20 p.m. – 2:40 p.m. 

• Session 2A: Breeding & Phenotyping (Ballroom) — Cotton Chronology: Convolutional Neural Network Enables Single-Plant Senescence Scoring with Temporal Drone Images.
  Aaron DeSalvio
• Session 2B: Livestock (Corps) — Computer vision optimization for smart beef cattle feed scoring in Calan gate systems.
  Egleu Mendes
• Session 2C: Education & Outreach (Ross) — Using ChatGPT with Novice Arduino Programmers: Effects on Performance, Interest, Self-Efficacy, and Programming Ability.
  Donald Johnson
• Session 2D: Horticulture – Field (Reveille) — "Comparing YOLOv8 and Mask RCNN for object segmentation in complex orchard environments."
  Ranjan Sapkota
• Session 2E: Management Cotton (Traditions) — An AI-Driven DigitalTwin Framework for Cotton Feature Forecasting and Yield Prediction.
  Pankaj Pal

2:40 p.m. – 3:00 p.m. 

• Session 2A: Breeding & Phenotyping (Ballroom) — High Throughput Phenotyping of the Energy Cane Crop Using UAS LiDAR Data.
  Benjamin Ghansah
• Session 2B: Livestock (Corps) — Automated Sow Posture Detection and Body Condition Estimation by 3D Computer Vision towards Precision Health Monitoring.
  Yibin Wang
• Session 2C: Education & Outreach (Ross) — Expanding AI expertise in the USDA’s Agricultural Research Service by connecting graduate students with researchers in a novel internship program.
  Brian Stucky
• Session 2D: Horticulture – Field (Reveille) — AI in the orchard: Improving sustainability through predictive yield in trees.
  Carolina Trentin
• Session 2E: Management Cotton (Traditions) — Digital Twin Model for In-Season Management, Biomass and Yield Forecasting of Cotton Crops.
  Juan Landivar

3:00 p.m. – 3:20 p.m.

• Session 2A: Breeding & Phenotyping (Ballroom) — Spatial Transformer Network YOLO Model for Plant Phenotypic Discovery.
  Yash Vivek Zambre
• Session 2B: Livestock (Corps) — Quantifying nesting behavior metrics of broiler breeder hens with computer vision and big data analytics.
  Aravind Mandiga
• Session 2C: Education & Outreach (Ross) — Knowledge Management in Agricultural Research through Esri’s ArcGIS Knowledge.
  Joseph Cullinan
• Session 2D: Horticulture – Field (Reveille) — Quantitative Analysis of Strawberry Runners for Breeding through Ground and Aerial Imaging.
  Xu Wang
• Session 2E: Management Cotton (Traditions) — In-season cotton yield prediction with scale-aware CNN models and UAV RGB imagery.
  Haoyu Niu

3:20 p.m. – 3:40 p.m.

• Session 2A: Breeding & Phenotyping (Ballroom) — Deep Learning and High-Throughput Phenotyping: Advancing Winter Wheat Breeding
  Swas Kaushal
• Session 2B: Livestock (Corps) — Classification and Regression Tree Approach for Prediction of E. coli Prevalence in Pasture Poultry Farms.
  Abhinav Mishra
• Session 2C: Education & Outreach (Ross) — Jaguza Tech developed an AI platform to provide digital assistance to the farmers both crops, livestock, and aquaculture working in Uganda and Nigeria Ibadan State.
  Katamba Ronald
• Session 2D: Horticulture – Field (Reveille) — Winter Damage Detection on Golf Courses through Drone-Based Multispectral Imaging.
  Ce Yang
• Session 2E: Management Cotton (Traditions) — Cotton Growth Forecast Using UAV Data with Bayesian Neural CDE.
  Lei Zhao

Panel 2 (Ballroom) 
3:45 p.m. – 5:30 p.m.

USDA/NSF AI Institutes Programs

• Dr. Soumik Sarkar, Iowa State (AIIRA)
• Dr. Vikram Adve, UIUC (AIFARMS)
• Dr. David Mulla, UMN (AI-Climate)
• Dr. Ananth Kalyanraman, WSU (AgAID)
• Dr. Ilias Tagkopoulos, UC Davis (AIFS)

Poster Session — Social & hors d’oeuvres (Foyer)
Graduate Students
5:30 p.m. – 7:00 p.m. 

Wednesday, April 17th

Breakfast (Foyer)
7:30 a.m. – 8:30 a.m.

Panel 3 (Ballroom)
8:30 a.m. – 9:45 a.m.

Water and AI

• Dr. Rabi Mohtar
• Dr. Susan O'Shaughnessy
• Dr. Joseph Quansah
• Dr. David Mulla 

Coffee Break (Foyer)
9:45 a.m. – 10:00 a.m.

Five Concurrent Breakout Sessions

10:00 a.m. – 10:20 a.m.

• Session 3A: Data Methods (Ballroom) — Data to Science Engine (D2SE) - A Data-driven Open Science Ecosystem for Sustained Innovation
  Jinha Jung 
• Session 3B: Social Implications of AI (Corps) — Smart Sensors for Enhancing Agricultural Water Conservation in Texas High Plains
  Mukesh Kumar Mehla
• Session 3C: Horticulture – Greenhouse & Controlled Env. (Ross) — Multimodal Sensing for On-Plant Size and Weight Estimation of Greenhouse Strawberry.
  Al Bashir
• Session 3D: Rangeland, Forestry, & Ecology (Reveille) — Harnessing Innovative Artificial Intelligence and Climate-Smart Technologies in Rangeland Systems.
  Marcia Fernandes
• Session 3E: Management Systems (Traditions) — Digital Twin Models: Implications for Farm Management.
  Yuri Calil

10:20 a.m. – 10:40 a.m. 

• Session 3A: Data Methods (Ballroom) — Enhancing Agricultural Practices through Multi-Object Tracking and Segmentation of Homogeneous Objects
  Zahra Khademi
• Session 3B: Social Implications of AI (Corps) — Effective Transfer and Adoption of AI technologies: Lessons learned from other agricultural innovations.
  Rafael Landaverde
• Session 3C: Horticulture – Greenhouse & Controlled Env. (Ross) — Quality Index Measurement System for Tomatoes based on Self-Attention Convolutional Neural Networks and Channel Pruning and Quantization.
  Yaqoob Majeed
• Session 3D: Rangeland, Forestry, & Ecology (Reveille) — Improving Strategies for Multi-step Prediction of Time Series using Convolutional Neural Network: A Case Study in Aboveground Vegetation Biomass Forecasting.
  Efrain Noa Yarasca
• Session 3E: Management – Systems (Traditions) — Web-based tool for early season stand counts using CNN on UAS imagery.
  Mahmoud Eldefrawy

10:40 a.m. – 11:00 a.m.

• Session 3A: Data Methods (Ballroom) — Advancing the Symbiosis and Synergy of Research and Extension: Transforming Food Systems’ Efficiency and Sustainability with Predictive AI Dissemination
  Chin-Ling Lee
• Session 3B: Social Implications of AI (Corps) — Factors Affecting Farmer Adoption of Unmanned Aerial Vehicles: Current and Future
  Tong Wang
• Session 3C: Horticulture – Greenhouse & Controlled Env. (Ross) — Scene Graph Generation from Point Cloud Data of Tomato Plants: Segmentation and Spatial Relationships.
  Prasad Nethala
• Session 3D: Rangeland, Forestry, & Ecology (Reveille) — Application of Remote Sensing and Artificial Intelligence for Detection and Quantification of Invasive Plant Encroachment in West Texas Rangelands.
  Sayantan Sarkar
• Session 3E: Management Systems (Traditions) — AgSkySight: Automating the UAV Image Data Processing Workflow for Precision Agriculture.
  Matheus Siqueira de Souza

11:00 a.m. – 11:20 a.m.

• Session 3A: Data Methods (Ballroom) — Application of TensorFlow model for identification of herbaceous mimosa (Mimosa strigillosa) from digital images
  Tri Setiyono
• Session 3B: Social Implications of AI (Corps) — Industrial Perspective of data rights and obligations in agriculture reflected by the key factors of agriculture technology providers
  Ziwen Yu
• Session 3C: Horticulture – Greenhouse & Controlled Env. (Ross) — Bruised Fruit Detection Using Deep Machine Learning Algorithm.
  Ping Zhang
• Session 3D: Rangeland, Forestry, & Ecology (Reveille) — "Proximal sensors provide a promising solution for obtaining reliable ground-based measurements of pasture height and biomass."
  Rashmi Sharma
• Session 3E: Management Systems (Traditions) — Machine Learning Algorithms for Fertilizer Application and Corn Cob Quality Prediction Based on Soil Nutrient Data.
  Binita Thapa

11:20 a.m. – 11:40 a.m.

• Session 3A: Data Methods (Ballroom) — Confident machine learning improves the prediction accuracy of ResNet-18
  Zhanyou Xu
• Session 3B: Toward sustainable culture media: Using artificial intelligence to optimize reduced-serum formulations for cultivated meat
  Reza Ovissipour
• Session 3C: Horticulture – Greenhouse & Controlled Env. (Ross) — Enhancing Crop Health: An Embedded Edge AI Solution for Real-Time Disease Detection
  Mike Ojo
• Session 3D: Rangeland, Forestry, & Ecology (Reveille) — Enhanced Image Classification of Agricultural Forage Plants and Weeds through a CNN Model Utilizing RMSprop Optimization.
  Aftab Siddique
• Session 3E: Management Systems (Traditions): Machine Learning for Predicting CO2 Emissions — Enhancing Climate-Smart Agricultural Systems with Biochar and Organic Amendments.
  Anoop Valiya Veettil

11:40 a.m. – 12:00 a.m.

• Session 3A: Data Methods (Ballroom) — TBD
• Session 3B: Social Implications of AI (Corps) — TBD
• Session 3C: Horticulture – Greenhouse & Controlled Env. (Ross) — TBD
• Session 3D: Rangeland, Forestry, & Ecology (Reveille) — Automating Forest Stand Delineation using AI Integration of Optical Imagery, Airborne LiDAR, and Forest Inventory Data.
  Can Vatandaslar
• Session 3E: Management Systems (Traditions) — Large-scale Climate Impacts on Midlatitude Rainfall Erosivity Patterns over
  the Contiguous United States Using Composite-Harmonic and AI-driven Cluster Analyses
  Jai Hong Lee

Lunch (Ballroom)
12:00 p.m. – 1:00 p.m. 

Keynote Speaker (Ballroom)
1:00 p.m. – 2:00 p.m.
Jennifer Clarke

Statistics and AI: Occam vs. Hickam

In this presentation, we will discuss two different philosophies behind scientific modeling that emphasize either simplicity or complexity, respectively. This tension underlies the approach of statisticians to predictive modeling. We will discuss how this tension is handled through a statistical lens with an example from regularization in regression contexts. We will then present how deep learning methods appear to take a different approach, explicitly ignoring this tension with (surprisingly) good results. We follow this with a hypothesis that may explain these results and the associated costs of modeling in high dimensions. We conclude with some challenges to the adoption and use of AI for modeling, with references for further study.

Coffee Break (Foyer)
2:00 p.m. – 2:30 p.m. 

Panel 4 - Industry/Technology (Ballroom)
2:30 p.m. – 3:45 p.m.

TAMU Phenotyping Greenhouse Tour Using AI Ethically Next Steps & Planning Optional Business Meetings (will run concurrently with the panel)
2:30 p.m. - 3:30 p.m.
Bus to Phenotyping Greenhouse, meet in pre-conference space (5-minute ride).

Using AI Ethically (Ballroom)
3:45 p.m. - 5:00 p.m.

Next Steps & Planning (Ballroom)
5:00 p.m. – 6:00 p.m.

Optional Business Meetings (Ballroom)
6:00 p.m. – 7:00 p.m.
 

NOTE: The agenda will be finalized closer to the event kick-off. *