As automated recording sensors (like cameras and microphones) become more powerful and less expensive, they have great utility for natural resource inventory and monitoring purposes. Sensors can provide information at frequent intervals at many survey sites, including those where human travel is difficult or dangerous. This greatly enhances our ability to determine the status and trends for a vast array of species in our parks. However, the data streams from those sensors (e.g. images, video, sound recordings) require processing to extract ecologically meaningful information. At the same time that the sheer volume and scale of the data streams we obtain with these sensors provides enormous value to species investigations, it also makes this work unwieldy and costly in terms of processing time and effort, essentially precluding traditional manual processing of data. Artificial Intelligence (AI) and Machine Learning (including computer vision, deep learning, etc.) are computer algorithms that are being applied to an ever-growing number of human and natural resource management challenges. Artificial Intelligence involves pattern detection. Common applications of AI in natural resources management include species identification, animal behavior classification, and biodiversity estimation in large datasets of remotely
recorded images, audio recordings, and videos. The NPS Inventory & Monitoring Division (IMD) is contending with ever-expanding datasets and the directive to report on those datasets in a timely manner. IMD is now beginning to evaluate AI and Machine Learning
needs for species inventories and monitoring across the 32 Inventory and Monitoring (I&M) networks.
This SIP position would be an integral contributor to understanding the current dispersed IMD AI efforts and how those could be coordinated into tools that may be used service-wide. The position may be structured in several possible ways depending on the specific skillset and interests of the SIP intern, including:
(1) Planning – The intern would assist IMD in evaluating what existing protocols and methodologies are currently being employed across the 32 I&M networks and parks (developing lists of instruments being used, type of data being collected, identifying resource management questions being informed, and requesting example data). The intern will then help to assess which protocols and projects have clear compatibilities to apply AI approaches, and develop flowcharts to help best develop plans to support these project needs with AI in the most efficient manner.
(2) Case Study – The intern could engage in a hands-on project to establish a “proof of concept” in applying AI to image or audio data. This project would directly benefit a group of networks with similar data, but would produce a ‘pipeline’ / SOP for processing similar
data from any source.
(3) Training Dataset – The intern could engage in collecting and refining image and/or audio data to develop high quality training datasets that will help in developing AI approaches. The paucity of useful training datasets for bioacoustics is a considerable hurdle in applying AI approaches. Developing training data would provide considerable benefits over the long-term by enabling other AI researchers to test their methods with our datasets, and thus enable direct applications of more advanced approaches.
This internship represents an integrated, interdisciplinary approach to aiding NPS to process natural resource data and address questions of pressing conservation concern. In effect, this internship is the first step in helping bring IMD into the modern age in tackling sensor (“Internet of Things”) data on species and ecosystems. What is started in this internship is just the beginning of more far-reaching application of these AI approaches, and will help NPS better identify, understand, and track the condition of its resources over the long-term.
This position is offered through the National Park Service's Scientists in Parks (SIP) Program in partnership with Stewards Individual Placement Program (Stewards) and The Geological Society of America (GSA). To apply, please visit https://rock.geosociety.org/eo/viewJob.asp?jobID=3129. To learn more about the NPS Inventory & Monitoring Program, please visit https://www.nps.gov/im/index.htm.
The intern will need to have a strong background in computer programming, machine learning, and artificial intelligence. The intern should be able to use Python, R, or other appropriate software to accomplish the goal. The intern should have some familiarity with wildlife and a strong passion for nature conservation, but a degree in wildlife ecology or natural resources is not required.
The applicant must be a U.S. citizen or U.S. permanent legal resident (“green-card-holder”) between the ages of 18 and 30 years old, inclusive, or veterans up to age 35. Prior to starting this position, a government security background clearance will be required.