Insurance for Pets

WWF in Indonesia uses machine learning to save orangutans – Insurance for Pets

A Sumatran orangutan mother with her baby. Image: WWF-Indonesia.

WWF-Indonesia’s mission is to save orangutans, which are critically endangered in Indonesia. Borneo orangutan populations have declined by more than 50% in the past 60 years, and the species’s habitat has been reduced by at least 55% in the past 20, according to the association. years.

The severe decline in the population of orangutans from Indonesia and Malaysia is directly linked to man and the cruel practices of poaching and illegal pet trade, as well as the destruction of their habitat .

As orangutans are generally solitary and spend a large part of their life in trees, it is difficult for associations to measure accurately the remaining populations, and each analysis or conservation action is complex and takes time. In order to speed up its efforts to save orangutans, WWF-Indonesia then turned to Amazon Web Services (AWS) to use machine learning services to better understand the size and health of populations of orangutans in their natural habitat.

Technology to the rescue of nature

The non-profit organization explains that this operation allows them to study more territories with fewer resources, and therefore to reduce operating expenses and to mobilize a larger part of the conservation funds to protect biodiversity in Indonesia.

« With careful use of technology, this innovation will help biologists and conservationists effectively and inexpensively monitor the behavior of wildlife over time. » This will allow us to allocate our resources to intensify monitoring efforts and invest more in conservation actions, ”explains Aria Nagasastra, finance and technology director of WWF-Indonesia.

“Collaboration between WWF-Indonesia and AWS on innovative technological solutions can help elevate biodiversity conservation practices in Indonesia to a higher level. « 

A male Sumatran orangutan. Image: WWF-Indonesia.

Faster analyzes …

For the past fifteen years, WWF-Indonesia has been evaluating the health of orangutan populations and preserving their 568,700 hectare habitat in Sebangau National Park, in central Kalimantan, Indonesia.

As the organization explained, these assessments require local experts and volunteers, who must go to the field daily to find the orangutans. They must then photograph them, then download the images to the computers of the camp and finally bring the data to the city so that they can be analyzed by a WWF expert.

To analyze each batch which contains thousands of photos, the teams have three days. The analysis time has now been reduced to less than 10 minutes.

… and more precise

Thanks to the Cloud, WWF-Indonesia now automatically collects images from mobile phones and motion-activated cameras at its base camp and uploads them to Amazon Simple Storage Service (Amazon S3) for analysis.

Through the use of Amazon SageMaker, data scientists and developers are able to quickly build, train and deploy machine learning models to scale.

Not only does the new process save time, but the organization also said it improves the accuracy and specificity of the data. These data include measures such as the gender ratio and age categories, as well as assessments of population viability. WWF-Indonesia can now quickly identify a pregnancy, illness or injury that requires immediate treatment.