▸Currently working with the AVERT team to improve the VIGIA-PlumNet model for the Cleveland Volcano in the Aleutian island chains. This volcano is closely monitored for potential eruption, since it is on the flight path from the US to Asia. Using a neural-net to determine non-ash plume emissions, we can automate this monitoring by detecting continuous release of gasses, which greatly decreases the likelihood of eruption. So far we have increased the class accuracy from 3% to 22% and plan to expand detection capabilities to include infrared camera for nighttime detection.