EPA Artificial Intelligence Inventory

EPA keeps an inventory of our AI projects in compliance with EO 13960 Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government.

Artificial Intelligence Project	Short Description of Project
Use of random forest model to predict exposure pathways	Ranking the potential risk posed to human health by chemicals requires tools that can estimate exposure from limited information. In this study, chemical structure and physicochemical properties were used to predict the odds that a chemical might be associated with any of four exposure pathways leading from sources-consumer (near-field), dietary, far-field industrial, and far-field pesticide-to the general population. The balanced accuracies of these source-based exposure pathway models range from 73 to 81%, with the error rate for detecting positive chemicals ranging from 17 to 36%. We then used exposure pathways to organize predictions from 13 different exposure models as well as other predictors of human intake rates. We created a consensus, meta-model using the Systematic Empirical Evaluation of Models framework in which the predictors of exposure were combined by pathway and weighted according to predictive ability for chemical intake rates inferred from human biomonitoring data for 114 chemicals. The consensus model yields an R2 of ∼0.8. We extrapolate to predict relevant pathway(s), median intake rate, and credible interval for 479 926 chemicals, mostly with minimal exposure information. This approach identifies 1880 chemicals for which the median population intake rates may exceed 0.1 mg/kg bodyweight/day, while there is 95% confidence that the median intake rate is below 1 μg/kg BW/day for 474572 compounds. Consensus Modeling of Median Chemical Intake for the U.S. Population Based on Predictions of Exposure Pathways
Records Categorization	The records management technology team is using machine learning to predict the retention schedule for records. The machine learning model will be merged into a records management application to help users apply retention schedules when they submit new records.
Enforcement Targeting	EPA’s Office of Compliance, in partnership with the University of Chicago, built a proof-of-concept to improve enforcement of environmental regulations through facility inspections by the EPA and state partners. The resulting predictive analytics showed a 47% improvement of detecting violations of the Resource Conservation and Recovery Act.

Artificial Intelligence Project

Short Description of Project

Use of random forest model to predict exposure pathways

Ranking the potential risk posed to human health by chemicals requires tools that can estimate exposure from limited information. In this study, chemical structure and physicochemical properties were used to predict the odds that a chemical might be associated with any of four exposure pathways leading from sources-consumer (near-field), dietary, far-field industrial, and far-field pesticide-to the general population. The balanced accuracies of these source-based exposure pathway models range from 73 to 81%, with the error rate for detecting positive chemicals ranging from 17 to 36%. We then used exposure pathways to organize predictions from 13 different exposure models as well as other predictors of human intake rates. We created a consensus, meta-model using the Systematic Empirical Evaluation of Models framework in which the predictors of exposure were combined by pathway and weighted according to predictive ability for chemical intake rates inferred from human biomonitoring data for 114 chemicals. The consensus model yields an R2 of ∼0.8. We extrapolate to predict relevant pathway(s), median intake rate, and credible interval for 479 926 chemicals, mostly with minimal exposure information. This approach identifies 1880 chemicals for which the median population intake rates may exceed 0.1 mg/kg bodyweight/day, while there is 95% confidence that the median intake rate is below 1 μg/kg BW/day for 474572 compounds.

Consensus Modeling of Median Chemical Intake for the U.S. Population Based on Predictions of Exposure Pathways

Records Categorization

The records management technology team is using machine learning to predict the retention schedule for records. The machine learning model will be merged into a records management application to help users apply retention schedules when they submit new records.

Enforcement Targeting

EPA’s Office of Compliance, in partnership with the University of Chicago, built a proof-of-concept to improve enforcement of environmental regulations through facility inspections by the EPA and state partners. The resulting predictive analytics showed a 47% improvement of detecting violations of the Resource Conservation and Recovery Act.

EPA Artificial Intelligence Inventory

Data