When an earthquake hits, Watson solution helps schools cope
For the 2018 Call for Code Global Challenge, a team of IBMers developed Frida, an AI and IoT solution to help teachers, students & emergency…
Born and raised in Quito, Ecuador, Sebastian Alvarado experienced several earthquakes growing up — both at home and in school.
“The one in my school was messier,” the intern Watson Studio software developer said. “There were so many people. Teachers were worried about students, and everyone started running everywhere.”
These childhood experiences drew Alvarado to the 2018 Global Call for Code Challenge and inspired the project “Frida.” This end-to-end solution predicts the magnitude of earthquakes based on sensor data and identifies the best escape routes and detects people trapped in damaged classrooms.
Today, IBM announced Team Frida as the winner of the internal Call for Code Challenge. The external winners will be announced at the Call for Code Global Prize Celebration on Oct. 29 at the Regency Ballroom in San Francisco.
“Frida mitigates natural disasters by combining emergency data with AI technology using IBM IoT platform, Watson Studio, and Watson Services,” said Lin Ju, Watson Studio senior development manager at IBM Canada Lab who led the team. “For our proof of concept, we focused on earthquakes in schools, but this solution can be applied to other areas.”
Frida consists of two offerings – a mobile AI-enabled application built through IBM Watson Studio and an IoT device with a Gyroscope sensor, a heat sensor, and a camera. Between the two offerings is the IBM Cloud, explained team member Simar Singh, a Watson Studio software developer.
“The IoT device streams data to the IBM Cloud, and Watson Studio uses that data and scores the model to predict if there’s an earthquake and its magnitude,” Singh said. “Later it’s pushed out to the app as an alert.”
To predict the earthquake’s magnitude, Frida leverages historical data from the U.S. Geological Survey and uses it to train deep learning models. These models are tested using the close-to-real-time data from the Incorporated Research Institutions for Seismology.
The result is accurate and timely alerts sent through the Frida mobile app to students, teachers, and first responders. All the data from the Frida IoT device is fed into a continuous learning loop to improve the model’s predictions.
“Our team is confident that these IBM technologies allow us to build a smarter and safer future,” Ju said. “The Call for Code Global initiative is a great opportunity to show the world that our developers can use the latest AI technology to have a real impact on society.”
That’s why IBM Canada Lab was so involved in the Call for Code, hosting a hackathon, and holding an internal showcase to demo all the submissions by lab teams, says Director Steven Astorino.
“Team Frida’s Call for Code submission displayed the ingenuity, drive and innovation we foster here at IBM Canada Lab,” Astorino said. “Lin, Simar, Sebastian and Lauren [Hartman] worked very hard on a cause that was close to their hearts. I could not be more impressed with their win.”
Alvarado is especially eager to bring the technology back home. Thinking back to an earthquake that hit Ecuador two years ago, he realizes that awareness of earthquakes is key.
“They didn’t know when [the earthquake] was going to come, and there were a lot of aftershocks,” he said. “If you could have told them about the aftershocks and to leave where they were, they could have been saved.”