WWF takes on machine learning for orang utan conservation
The World Wildlife Fund for Nature (WWF), Indonesia
is using machine learning to protect some of most endangered species in Southeast Asia.
Edwin Chaidir, Manager of Information Technology, WWF shared that WWF is studying the orang utan population in Borneo, which is losing 2,000 orang utans a year due to factors such as habitat loss, illegal trade, and forest fires.
"The future of this population depends on what we do now and what we do in the future," he said.
He said the critically-endangered creatures are hard to identify in the wild, so WWF needs the help of human experts when it surveys the orang utan population. An AWS machine-learning prototype, WWF Orangutan Face Recognition v0.1, has now automated the process, monitoring the population and identifying individual orang utans in habitats in Kalimantan, Borneo.
This decreases dependencies on local experts and allows the WWF to invest more on resources to protect the wildlife, Chadir said.
AWS created the model
based on physical characteristics
of orang utans. Several services were used, including
S3, SageMaker, Lambda and API Gateway:
- S3 allows users to store and retrieve data from anywhere as it is stored in the AWS cloud.
- SageMaker is a fully-managed service can build, train, and deploy machine learning models quickly.
- Lambda is a service that runs code in response to event triggers, while automatically managing the underlying compute resources. In the past, anyone who wanted to create such a capability in the cloud would have had to specify exactly what cloud computing resources would be needed.
- API Gateway is a fully-managed service that makes it easy for developers to create, publish, maintain, monitor, and secure APIs at any scale. APIs act as a bridge so that applications can access data, business logic, or functionality from elsewhere.
At the time of Chaidir's speech, accuracy ranged from 35% to 99%. The plan is to raise confidence levels to 75-99%, Chaidir said. With machine learning, systems tend to improve over time as they are exposed to more data.
"We believe that technology can overcome the challenges of field data collection. It can facilitate innovation and the creation of new ideas and tools," he said.
"It is only just the beginning," he added, noting that the technology can extend to identifying tigers, sharks, or help with the illegal wildlife trade. "Together we can protect the very things that keep us alive, because we believe that everything is possible," he said.
The WWF case study was shared at the first-ever AWS ASEAN Public Sector Summit in Singapore in September 2019.
Edwin Chaidir, Manager of Information Technology, WWF shared that WWF is studying the orang utan population in Borneo, which is losing 2,000 orang utans a year due to factors such as habitat loss, illegal trade, and forest fires.
"The future of this population depends on what we do now and what we do in the future," he said.
He said the critically-endangered creatures are hard to identify in the wild, so WWF needs the help of human experts when it surveys the orang utan population. An AWS machine-learning prototype, WWF Orangutan Face Recognition v0.1, has now automated the process, monitoring the population and identifying individual orang utans in habitats in Kalimantan, Borneo.
This decreases dependencies on local experts and allows the WWF to invest more on resources to protect the wildlife, Chadir said.
 |
| Add caption |
- S3 allows users to store and retrieve data from anywhere as it is stored in the AWS cloud.
- SageMaker is a fully-managed service can build, train, and deploy machine learning models quickly.
- Lambda is a service that runs code in response to event triggers, while automatically managing the underlying compute resources. In the past, anyone who wanted to create such a capability in the cloud would have had to specify exactly what cloud computing resources would be needed.
- API Gateway is a fully-managed service that makes it easy for developers to create, publish, maintain, monitor, and secure APIs at any scale. APIs act as a bridge so that applications can access data, business logic, or functionality from elsewhere.
At the time of Chaidir's speech, accuracy ranged from 35% to 99%. The plan is to raise confidence levels to 75-99%, Chaidir said. With machine learning, systems tend to improve over time as they are exposed to more data.
"We believe that technology can overcome the challenges of field data collection. It can facilitate innovation and the creation of new ideas and tools," he said.
"It is only just the beginning," he added, noting that the technology can extend to identifying tigers, sharks, or help with the illegal wildlife trade. "Together we can protect the very things that keep us alive, because we believe that everything is possible," he said.
The WWF case study was shared at the first-ever AWS ASEAN Public Sector Summit in Singapore in September 2019.
Comments
Post a Comment