A team of researchers is using artificial intelligence in hopes of saving U.S. ranchers $35 million spent keeping an invasive plant in check.
Researchers at Carnegie Mellon University (CMU) and a conservation ranch in Montana have been training AI models to detect an invasive species even when there is limited data.
Leafy spurge is a weed with small green flowers that can wreak havoc on the ecosystem. It is toxic to livestock and crowds out native plants. When it spreads, leafy spurge can render a whole hayfield inedible.
The plant causes an estimated $35 million annually in losses in beef and hay production. The weed also shrinks grazing land as it crowds out native plants. Birds and pollinators also lose habitats as the invasive species spreads, causing changes to the entire ecosystem.
Scientists hope the technology can help identify and monitor leafy spurge to help prevent it from spreading. They hope to use generative AI to improve existing models that are trained to detect leafy spurge using drone images.
They used a technique to get AI to create images of the invasive plant based on existing drone footage, creating training data that showed leafy spurge under various conditions, saving ecologists from having to go out to gather the data in different weather conditions or locations.
The hope is to expand the effort to manage other invasive species as well.