AIIRA - AI Institute For Resilient Agriculture

Recent paper in PNAS Nexus "InsectNet: Real-time identification of insects using an end-to-end machine learning pipeline." Congratulations to all authors, particularly the first author, Ph.D. Student Shivani Chiranjeevi and the corresponding authors Drs. Arti Singh and Baskar Ganapathysubramanian. Short summary: InsectNet is an innovative deep-learning approach for automated insect identification in agricultural settings. The system addresses a critical need in agriculture: accurately distinguishing between beneficial and harmful insects for better pest and insect management. Extended summary: Insect pests significantly impact global agricultural productivity and crop quality. Effective integrated pest management strategies require the identification of insects, including beneficial and harmful insects. Automated identification of insects under real-world conditions presents several challenges, including the need to handle intraspecies dissimilarity and interspecies similarity, life-cycle stages, camouflage, diverse imaging conditions, and variability in insect orientation. Key features of this work: (a) using a large dataset of insect images collected through citizen science along with label-free self-supervised learning to train a global model; (b) fine-tuning this global model using smaller, expert-verified regional datasets to create a local insect identification model; (c) provides high prediction accuracy even for species with small sample sizes, (d) designed to enhance model trustworthiness, and (e) democratizes access through streamlined machine learning operations (MLOps). This global-to-local model strategy of InsectNet offers a more scalable and economically viable solution for implementing advanced insect identification systems across diverse agricultural ecosystems. We report accurate identification (>96% accuracy) of numerous agriculturally and ecologically relevant insect species, including pollinators (e.g., butterflies, bees), parasitoids (e.g., wasps and flies), predators (e.g., lady beetles and mantises), and harmful insects (e.g., armyworms, cutworms, grasshoppers, stink bugs). InsectNet provides fine-grained insect species identification, works effectively in challenging backgrounds, and avoids making predictions when uncertain, increasing its utility and trustworthiness. The model and associated workflows are available through a web-based portal accessible through a computer or mobile device. We envision InsectNet complementing existing approaches and part of a growing suite of AI technologies for addressing agricultural challenges. AIIRA - AI Institute For Resilient Agriculture TrAC - Translational AI Center https://lnkd.in/gC2KWu_3

fortunate to be part of this great team...

Baskar Ganapathysubramanian and Chinmay Hegde highlight how AIIRA is using Llama to solve real-world problems. https://lnkd.in/gfWKik3g

Congratulations Sarah Jones

Heartiest congratulations to Liza Van der Laan, who successfully defended her Ph.D. thesis earlier this week Iowa State University. Liza is a superstar researcher with a long list of accolades. She has 13 research papers or reviews - published, accepted, in-review, submitted or near-submission. Liza has worked on several projects in the group, providing leadership in research in heat stress related topics. She has worked in the breeding program, leading small and large teams. She has shared her work at several national and international meetings and conferences. She led a grant writing initiative on abiotic stress, funding part of her research, through the Iowa Soybean Research Center. He has led a group of engineering students in their Capstone projects. She has led an interdisciplinary project in our team. She has served as a teaching assistant and lectured in AGRON5210 course on cultivar development and guest lectured in a Robotics and AI for Agriculture course at Carnegie Mellon University. Liza has been an active contributor and member of the department of Agronomy Iowa State University - College of Agriculture and Life Sciences. She was actively involved in the department graduate student club, Baker Plant Breeding Symposium, and Women in Ag and AI. She is a graduate of the ISU’s preparing future faculty program, and Emerging Leaders Academy. Liza has a long list of accolades and achievements, including the ISU Brown Graduate Fellowship, Agronomy Outstanding Ph.D. student award, Weber award for excellent in Plant Breeding, Corteva Agriscience New Frontiers Scholar and NAPB, National Association for Plant Breeding Borlaug Scholar. Thanks to funders: AIIRA - AI Institute For Resilient Agriculture Iowa Soybean Association Iowa Soybean Research Center TrAC - Translational AI Center

Congratulations to Sarah and Timi! Thank you for your exemplary contributions!

Most plant properties (phenotypes) are determined by a mix of genetics (G) and environment (E). But wouldn't it be simpler if some phenotypes were determined by genetics and others by environment? That's what Anirudha Powadi, Talukder Zaki Jubery & Michael Tross set out to do in this new preprint. They designed a new computational encoder strategy to decompose hyperspectral reflectance data from a maize diversity panel of ~700 genotypes grown in two environments into a set of genetic and environmental factors that can be used to predict whole plant traits. Anirudha Pawadi, Talukder Zaki Jubery, Michael Tross, James Schnable, Baskar Ganapathysubramanian (2024) Disentangling Genotype and Environment Specific Latent Features for Improved Trait Prediction using a Compositional Autoencoder. https://lnkd.in/gw-kzWHD Really fun project supported by AIIRA - AI Institute For Resilient Agriculture.

The women in Women in Ag and AI represent some of the best of bringing Ag and AI together. Thank you for bringing awareness to the next generation of change makers!

I was thrilled to witness the incredible advancements in agriculture research using cutting-edge, high-throughput technologies at MLCAS2024 and IPPS8. A big thank you to the organizers for creating such an inspiring platform! I am excited to have the opportunity to present a portion of my PhD work on automated phenotyping of mungbean pod traits. Leveraging AI and machine learning, this research aims to support mungbean breeding by enhancing precision and efficiency in identifying key traits for crop improvement. The future of agriculture is digital, and I'm proud to contribute to this transformative journey! #mungbean #iowastateuniversity #aiira #plantbreeding #MLCAS2024 #IPPS8

Happy to share that MLCAS 2024 (https://2024.mlcas.site/) was a great success! 150+ attendees, six great speakers, 30+ exciting contributed talks and posters, and an intriguing panel. Thanks to all participants, sponsors, and volunteers. Special thanks to International Plant Phenotyping Network - IPPN for making us a part of IPPS8. Let's keep the conversation going to realize the full potential of Cyber-Agricultural Systems. #ai, #machinelearning, #agriculture. TrAC - Translational AI Center, AIIRA - AI Institute For Resilient Agriculture, COALESCE, Iowa State University Plant Sciences Institute, John Deere, Corteva Agriscience, Bayer. Arti Singh, Baskar Ganapathysubramanian, Wei GUO, Asheesh Singh, James Schnable, Pat Schnable.