Artificial Intelligence and machine-learning technology are increasingly used in various sectors. Now, Google and Harvard University are working towards predicting yet another natural disaster if I may. According to a blog post by Google, the search engine giant has collaborated with Ivy league Harvard University where the technology is being developed to predict aftershock locations, something which is not new and has been attempted several times.
Earlier, a huge dump of data gathered from techniques used such as from Earth’s core or fiber optics, which means there is no shortage of availability of data to extrapolate. Therefore, Google and Harvard decided to collaborate and build an AI model that could possibly predict aftershocks from earthquakes for up to a year.
For the model, timing and size related data were mined using empirical laws along with the huge amount of data already available, however, forecasting the location of aftershocks is challenging. That’s where experts in machine learning from Google came into play. The collab team took data of 118 major earthquakes around the world and fed it onto the system based on neural network.
When the data was processed, researchers were able to make an AI model that could forecast aftershock locations. Although the technology is currently in its immature stage, it will soon become a major breakthrough in aftershock forecasts that could save lives as well as reduce the impact of the aftershocks worldwide. When the data was out, researchers involved in the study found that the AI model was able to locate physical quantities that would act as factors determining the role of the earthquake besides other data.
As of now, the study is published in the journal Nature. Researchers will have to work intensively on the technology to generate promising and reliable outcomes.