Saving the Elephants with AI
Ahead of this year’s World Elephant Day, Inspur joined with Xishuangbanna National Nature Reserve to develop an extensive technology system for the conservation of some 300 Asian elephants in Yunnan, China.
The Asian elephant, found only in South and Southeast Asia and the southern border of Yunnan, had been critically endangered for decades. 30 years of conservation brought Yunnan’s population of Asian elephants from a precarious 183 to 300, and even such a modest increase has already caused much conflict in the area due to shrinking habitats. The elephants frequently cross into human communities, causing costly disruptions like crop raids, structural damage, and danger to human lives. Moreover, increased interaction with humans can cause long-term changes to the animals’ behavior, which need to be monitored closely as well.
Most challenges to wildlife conservation are: sighting the animals at night, tracking them across vast and untraversable terrain, and the age-old problem that there simply is not enough manpower to efficiently do the work required.
A researcher tests his video recording device before installing it in the forest. A broad network of such devices all over Xishuangbanna National Nature Reserve capture image and footage of elephants to enable real-time tracking and analytics.
At the Xishuangbanna National Nature Reserve, Inspur installed a network of total solutions, from devices to powerful computational platforms to software systems, to aid the effort. Armed with more data and greater insights derived from advanced cloud computing, IoT, big data, and AI, researchers at the reserve are able to overcome logistical and resource challenges to implement strategic changes.
Increasingly, conservationists are learning that their work need not solely rely on ground teams of biologists, researchers and park rangers. Elsewhere in the world, other elephant researchers are finding an ally in AI.
Tracking Elephant Calls
Despite being the largest mammal on land, elephants can be remarkably hard to track, especially the forest elephants in Africa which live under dense rainforest canopies. One researcher at Cornell University decided to change his strategy and began collecting audio recordings in the forest to listen to the vocalizations of elephants. But he ended up with a whole new problem, which is that he had months of continuous recordings of the entire rainforest and no human way to parse all of that audio data. This was where AI came in: a neural network was trained to sift the data and distinguish elephant calls and communication patterns from ambient forest noise and sounds of other animals. Through deep learning, the Elephant Listening Project was not only getting an accurate, ongoing count of the elephants, but deriving other actionable information as well, such as the movement patterns of the animals at certain times in the year.
Tracking Asian elephant movement in Yunnan using a network of devices, data analytics and a visualization system
A further development of the Elephant Listening Project was that conservationists began to use it to listen for the gunshots of poachers. By finding out when and where the gunshots were going off, they could build a predictive model to prevent the poachers before they attack. Elsewhere, park rangers at Liwonde National Park are working on a similar goal. First, they used AI and predictive analytics to restore and manage a wildlife preserve on the brink of collapse. Then they developed a real-time visualization system that helps rangers stay on top of incidents in the park, and efficiently allocate resources and plan their patrols. They were also able gain insights from weather, animal movement and even sociocultural patterns to predict when and where poachers might strike. Just one example of the system’s effectiveness: the sophisticated network of sensors and AI algorithms had helped the rangers at Liwonde apprehend a notorious poacher in the area, responsible for multiple offenses.
Each industrial revolution has largely been a boon for humanity but often at the cost of the environment. But the latest industrial IT revolution can be an ally to, not an enemy of, environmental equity and ecological stewardship. It is the human will, compassion and innovation that drive our collective vision for a better planet, but the limitless computational capabilities of artificial intelligence can help make that vision a reality.