Click here to download PDF copy of presentation

Speaker: Julien Martin (US Geological Survey)

Abstract:

Applications of artificial intelligence (AI) and Machine learning (ML) techniques to ecological research and management are growing rapidly. These techniques have the potential to increase the efficiency of data collection and analyses of ecological data to help inform management decisions. In this session, we will discuss the benefits and limitations of these methodologies for natural resource management. More specifically, we will examine the roles and limitations of AI and ML in the context of structured decision making. Structured decision making (SDM) for natural resource management is a method for analyzing a decision by decomposing the problem into components, which include: (1) management objectives; (2) potential management actions; (3) models to project the consequences of actions; (4) optimization methods and (4) monitoring. We will talk about the benefits that AI/ML can bring in the context of SDM, with a particular focus on the following components: models, optimization and monitoring. Examples of successful applications of AI/ML include image analysis and the processing of acoustic data to identify species or matching individuals. In turn this identification process can be used to model vital rates, abundance and distribution of organisms. The accounting of imperfect detection and misclassification errors is important for reliable inference from data collected or processed with AI/ML. We will describe SDM and adaptive management cases studies that could benefits from AI/ML techniques. Finally, we will discuss the challenges and limitations of applying AI/ML to natural resource management.