AI for Conservation refers to the application of artificial intelligence to conservation, such as wildlife protection and the conservation and responsible stewardship of natural resources. For example, in the green security domain, the repeated and strategic interaction between those who protect these resources and those who seek to attack or exploit these resources can be modeled using game theory as a repeated game. While our predictive analytics effort focuses on predicting where adversaries (e.g., poachers) will strike, our prescriptive analytics work provides recommendations to defenders (e.g., rangers) to conduct strategic, randomized patrols. These analytics can be supported using machine learning, for example by detecting poachers or animals in unmanned aerial vehicle (UAV) imagery automatically.
Project Spotlight: PAWS - Protection Assistant for Wildlife Security
The PAWS AI system has been developed by CRCS researchers and collaborators. The Protection Assistant for Wildlife Security (PAWS) is an artificial intelligence (AI) software for wildlife protection. PAWS has been deployed in collaboration with wildlife conservation agencies to assist rangers around the world in protecting endangered wildlife.
PAWS helps conservation managers plan more informed patrol strategies to combat wildlife poaching, illegal logging, and illegal fishing. Taking in past poaching records and data about the geography of the protected area, PAWS uses machine learning to predict poachers’ behavior. After learning a model of poaching behavior, PAWS produces poaching risk maps and suggests patrol routes for rangers.
PAWS led to removals of thousands of traps used to kill and maim endangered wildlife in national parks in countries such Cambodia and Uganda. Furthermore, PAWS is integrated with the SMART software, which is a monitoring and reporting tool used in over 800 protected areas around the world. This means that PAWS is available for use at hundreds of national parks across the globe. The research areas we focus on here merges game theory and machine learning, in an approach called Green Security Games.
Multi-armed bandits to balance exploration and exploitation (AAAI'21)
Our paper Dual-Mandate Patrols: Multi-Armed Bandits for Green Security, which was selected as a Best Paper Runner Up at AAAI 2021, focuses on helping rangers choose where to patrol in a protected area, balancing exploration of infrequently visited regions and exploitation of known hotspots. Our proposed LIZARD algorithm bridges the gap between combinatorial and Lipschitz bandits, presenting an approach that achieves theoretical no-regret and performs well in experiments on real-world poaching data.
For more information about PAWS, please visit the Teamcore PAWS webpage.