Dramatically economize the monitoring for target pests and modernize how the USDA conducts wood borer surveys

The University of Florida's Division of Sponsored Research received a $147,186 cooperative agreement from the U.S. Department of Agriculture's Animal and Plant Health Inspection Service (APHIS) under the Plant and Animal Disease, Pest Control, and Animal Care program (CFDA 10.025), awarded September 30, 2025, with completion targeted for September 29, 2026. The Forest Entomology Lab will develop an artificial intelligence-based invasive bark beetle screening tool designed to modernize USDA wood borer survey operations and optimize pest monitoring workflows for CAPS (Cooperative Agricultural Pest Survey) identifiers and the National Identification Services. The AI classifier will be trained on domestic bark beetle species prevalent in the Eastern United States, common exotic species already established in the region, and high-priority exotic species not yet present in the country, with integrated functionality to identify and flag previously unknown species encountered during surveys.

The primary deliverable is an operational AI-based identification system capable of accurately distinguishing regulatory-concern bark and ambrosia beetles from Asia and Europe while minimizing false positives through training on non-regulatory species. Immediate beneficiaries include APHIS staff responsible for processing large volumes of beetle specimens collected during pest surveys, with broader benefits extending to tree-related industries and environmental conservation stakeholders. The award includes no subawards and culminates in a collaborative workshop in Raleigh, North Carolina, where the development team and agency staff will discuss machine learning applications for future survey protocols.

The university of florida forest entomology lab will develop an AI-based (artificial intelligence) invasive bark beetle screening tool for the workflow of the caps identifiers and the national identification services. The team will use their existing beetle identification model and train IT on all available domestic species and genera, primarily eastern US, on common exotic species established in the eastern US, and then on a set of high-priority exotics not yet in the country. Lastly, we will work together with the sponsor to organize a workshop in raleigh, NC that will enable the team and the agency staff to discuss the use of machine learning in future surveys. Deliverables and expected outcomes: what are the goals of the award and what evidence is required from the recipient? This might be research results, completion of a study or survey work, a completed report, etc. The goals and deliverables of the award include primarily an AI-based classifier for identification of exotic bark and ambrosia beetles of regulatory CON cern from asia and europe. The AI classifier will be trained on common bark beetles from the eastern US and of most common bark beetles that are not of regulatory concern, and then on the high priority pests from asia. A function to report the unknown unknowns w ILL be built in, for the model to recognize and report a new species that was not present in the training set. The immediate beneficiaries of the tool developed here are mainly agency staff responsible for the identification of the vast numbers of bark beetles generated in pest surveys. Ultimately, the beneficiaries include stakeholders in all tree-related industries and environmental conserva tion. This award has no subawards.