When we hear Artificial Intelligence we imagine a lot of different futuristic (or not so distant) technology. However, in the humanitarian sector there is, in fact, a wide range of applications for machine learning and highly accurate data prediction that can represent the next steps for disaster relief.

The Netherlands Red Cross is developing 510, a new initiative in the Philippines for better aid response after a natural disaster strikes. This initiative combines key factors such as wind speeds, rainfall amount in affected areas, and collected patterns from past disasters. All of this factors is pulled together in a Priority Index that can be deployed within hours after an event to identify affected areas and people.  

A real test for the initiative 

This was the house where Baslita used to live.

This was the house where Baslita used to live.

In October 2016, Typhoon Haima destroyed and damaged more than 198,000 homes and took at least 14 lives (AFP). Identifying affected peoples can often take several weeks, as it is a difficult and time consuming process that requires a lot of coordination and consideration of security factors. The first Priority Index was released within 24 hours with highly-accurate prediction of affected communities. This contributed to a faster supply of aid distribution to the worst-hit communities.

This technology was proven to be accurate, after a comparison with data on damaged property, which was provided by the Department of Social Welfare and Development (DSWD) and the National Disaster Risk Reduction and Management Council (NDRRMC).

What’s next?

This machine-learning method is intended to be replied to different countries, feeding from local data and adjusting to each community context. The research team is still working to make damage prediction more accurate. There are some blind spots to cover related to disaster scale. In order to tackle this problem, more testing and data collection is required, but the initiative, until now, has been positive.

This system is also being tested in different disaster response environments – for example in the earthquake zone, in Nepal and Africa.

Do not miss the complete story of 510 here.

Pin It on Pinterest

Shares
Skip to toolbar