Saturday, April 23, 2016

Outwitting poachers with AI

A century ago, more than 60,000 tigers roamed the wild. Today, the worldwide estimate has dwindled to around 3,200. Poaching is one of the main drivers of this precipitous drop. Whether killed for skins, medicine or trophy hunting, humans have pushed tigers to near-extinction. The same applies to other large animal species like elephants and rhinoceros that play unique and crucial roles in the ecosystems where they live.

Human patrols serve as the most direct form of protection of endangered animals, especially in large national parks. However, protection agencies have limited resources for patrols.

With support from the National Science Foundation (NSF) and the Army Research Office, researchers are using artificial intelligence (AI) and game theory to solve poaching, illegal logging and other problems worldwide, in collaboration with researchers and conservationists in the U.S., Singapore, Netherlands and Malaysia.

“In most parks, ranger patrols are poorly planned, reactive rather than pro-active, and habitual,” according to Fei Fang, a Ph.D. candidate in the computer science department at the University of Southern California (USC).

Fang is part of an NSF-funded team at USC led by Milind Tambe, professor of computer science and industrial and systems engineering and director of the Teamcore Research Group on Agents and Multiagent Systems.

Their research builds on the idea of “green security games” — the application of game theory to wildlife protection. Game theory uses mathematical and computer models of conflict and cooperation between rational decision-makers to predict the behavior of adversaries and plan optimal approaches for containment. The Coast Guard and Transportation Security Administration have used similar methods developed by Tambe and others to protect airports and waterways.

“This research is a step in demonstrating that AI can have a really significant positive impact on society and allow us to assist humanity in solving some of the major challenges we face,” Tambe said.

PAWS puts the claws in anti-poaching

The team presented papers describing how they use their methods to improve the success of human patrols around the world at the AAAI Conference on Artificial Intelligence in February.

The researchers first created an AI-driven application called PAWS (Protection Assistant for Wildlife Security) in 2013 and tested the application in Uganda and Malaysia in 2014. Pilot implementations of PAWS revealed some limitations, but also led to significant improvements.

PAWS uses data on past patrols and evidence of poaching. As it receives more data, the system “learns” and improves its patrol planning. Already, the system has led to more observations of poacher activities per kilometer.

Its key technical advance lies in its ability to incorporate complex terrain information, including the topography of protected areas. That results in practical patrol routes that minimize elevation changes, saving time and energy. Moreover, the system can also take into account the natural transit paths that have the most animal traffic – and thus the most poaching – creating a “street map” for patrols.
“We need to provide actual patrol routes that can be practically followed,” Fang said. “These routes need to go back to a base camp and the patrols can’t be too long. We list all possible patrol routes and then determine which is most effective.”

The application also randomizes patrols to avoid falling into predictable patterns. “If the poachers observe that patrols go to some areas more often than others, then the poachers place their snares elsewhere,” Fang said.

Since 2015, two non-governmental organizations, Panthera and Rimbat, have used PAWS to protect forests in Malaysia. The research won the Innovative Applications of Artificial Intelligence award for deployed application, as one of the best AI applications with measurable benefits.

The team recently combined PAWS with a new tool called CAPTURE (Comprehensive Anti-Poaching Tool with Temporal and Observation Uncertainty Reasoning) that predicts attacking probability even more accurately.

In addition to helping patrols find poachers, the tools may assist them with intercepting trafficked wildlife products and other high-risk cargo, adding another layer to wildlife protection. The researchers are in conversations with wildlife authorities in Uganda to deploy the system later this year. They will present their findings at the 15th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2016) in May. “There is an urgent need to protect the natural resources and wildlife on our beautiful planet, and we computer scientists can help in various ways,” Fang said. “Our work on PAWS addresses one facet of the problem, improving the efficiency of patrols to combat poaching.”


No comments: