Machine learning methods could more than double the number of environmental violations detected, says new research. The finding could prevent some of the worst effects of hurricanes.
When Hurricane Florence hit North Carolina, it caused dangerous bacteria and heavy metals to be washed into nearby waterways. Because state and federal environmental regulators are often underfunded, the worst effects could not be prevented.
However, machine learning could provide the solution to this issue, according to new research conducted by Stanford University scientists. Computers trained to automatically detect patterns in data could in future catch two to seven times as many infractions as current approaches.
“Especially in an era of decreasing budgets, identifying cost-effective ways to protect public health and the environment is critical,” commented study co-author Elinor Benami in a statement.
Machine learning methods could help predict where funds can yield the most benefit, offering support to overextended environmental regulators. The new research, which focused on the Clean Water Act, found that the authorities responsible for regulating more than 300,000 facilities in the USA inspect less than 10 per cent of those in a year.
Despite the potential of machine learning, the method also has drawbacks, warned the researchers in the statement. For example, farm owners may alter their behavior if they know their likelihood of being selected by the algorithm, while institutional, political and financial constraints could limit machine learning’s ability to improve upon existing practices.
“This model is a starting point that could be augmented with greater detail on the costs and benefits of different inspections, violations and enforcement responses,” concluded co-author Nina Brooks.
Photo credit: Waterkeeper Alliance Inc./ CC BY-NC-ND 2.0