by Gitanjali Rao
Grade 6 - Brentwood Middle School (Brentwood, TN, United States)
Third Place
Deep in the Virunga volcanic mountains, in the Congo basin, every month, approximately, 40 mountain gorillas are killed. Once these majestic creatures roamed the forests in the high volcanic mountains of Uganda, Rwanda and Congo in thousands, but today there are fewer than nine-hundred mountain gorillas on the planet. There is a high urgency to conserve these gorillas. I am particularly interested in the conservation of mountain gorillas because they are an important part of the Congo basin mountain forest ecology, as well as a critical link in the human evolution. Being the second closest relative to humans in terms of genetic makeup, their study can offer immense clues on human origins and, possibly cures for human diseases. The mountain gorillas face several threats today ranging from loss of habitat, to poaching, and spread of diseases. It does not help that they are in region that is in midst of war.
Today, several conservation efforts are underway to save the gorilla from extinction. These include increased security and patrolling, larger protected sanctuaries, and public education. There is also extensive use of technology such as motion sensors, cameras and GPS tracking. Besides these, one of the most important and emerging tool in the hands of conservationists is insights provided by processing data from the various sources. Lot of the conservation efforts today are decided based on data gathered through manual surveys, trackers, sensors, GPS and camera traps, which provide valuable details about gorilla movements, poacher activities and protected land incursions. Though these are effective to a certain extent, they are also very resource intensive, and manual which are prone to errors and biases. The data is also not always reliable or complete enough for actionable decisions. While data-driven approach is the best solution, there is a need for solution that can process large amounts of diverse data and provide accurate actionable results.
One technology that builds on the current efforts, but addresses some of the challenges, is the use of Artificial Intelligence (AI) for generating, something called predictive insights or analytics. Instead of relying on data processing abilities of humans and challenges with accuracy, AI specifically machine learning, can be used to process data to effectively model, and predict, conditions that are best suited for gorilla conservation. My interest in the idea came from reading about them in magazines and from my coding camps on data analysis, data science and AI. A machine learning engine perceives its environmental variables, and using techniques such as self-learning, predicts the probability of an event happening. In our case, this can help identify an event that is a potential threat to the gorillas. Machine learning is a smart way of writing programs, by allowing them to automatically learn from the data we provide and by correcting any mistakes. It thinks like a human, but stores, processes and predicts based on actual data, without bias, feelings or interpretation. I believe that predictive analysis built on a self-learning system can analyze huge amounts of data from many sources to create an accurate model of future behavior of scenarios such as, poacher movements, gorilla migration, disease spreads, and park land encroachment.
A similar idea is currently being tested is the PAWS* AI program for elephant conservation. However, PAWS is focused on poaching of African elephants. Mountain gorillas have a different set of environment data, including unique risks such as disease spread from humans that require the use of advanced machine learning capabilities to create a model with higher accuracy, for taking actions.
My idea of using machine learning and predictive analysis, builds on current traditional data analysis approaches, and determines the following four things:
- Rate of contraction of gorilla habitat and expansion of humans,
- The movement patterns of poachers,
- Behavior and migration patterns of gorillas,
- Tracking the spread of diseases within their habitat.
To achieve this, we would first need data to “seed” the system, also called training-data, from variety of sources ranging from weather patterns, past poacher movements, past gorilla tracking and geological land surveys. We can then use machine learning techniques to find relations of data, and eventually determine the likelihood of an event happening that hurts or favors the mountain gorilla. This can help the AI system understand and learn the unique aspects of a specific gorilla or poacher. The system can then suggest actions with better accuracy, which conservationists or rangers can execute. These can help with deployment of forest guards, disease control or habitat protection.
Machine learning models also have diverse uses. If we provide right data and the engine is “trained”, the model can be reused for other environmental conservation efforts with minimal work. Farmers, to plan their crops based on weather patterns, and, biologists, to study changing ecosystem can use the same concept.
Some constraints should be addressed in this solution. Gorillas reside in the intersection of three countries. Coordinating data collection activities, managing any privacy or data protection laws and coordinating actions against poachers based on a computer models, may not be ethical or lawful until an event occurs. Similarly, expansion of the habitat as a solution will require working within the forest and land laws of the several governments. However, we can make it happen with a collaborative effort and can mobilize active participation from people who are currently active in conservation, such as, field researchers, park rangers and biologists. We will also need a new group of people like engineers and data scientists who bring in special skills to manage and process data. We can implement gradually starting with a prototype , then a pilot to just address one scenario, and then test its effectiveness before expanding.
Substantial money and effort today goes into gorilla conservation, but lot of it is inefficient and wasteful expenditure. My proposal will have initial costs, especially for data collection, storage and processing, but there is lot of re-use. In addition, this kind of technology solution lends itself very well for crowd-sourcing. Data gathering and processing contributions can be opened up globally over the internet and we can easily find volunteers with skills who are willing to help.
Currently, the main organization involved in conservation of mountain gorillas is the International Gorilla Conservation Programme, which was established in 1991 as a joint effort between different wildlife foundations. I believe the same organizations can continue providing funds and resources to prioritize actions.
It is our responsibility as engineers to save the future of the planet and with advances in tools and techniques; we are guaranteed to have better results of conserving the mountain gorillas than ever.
References
*Protection Assistant for Wildlife Security “Artificial Intelligence and Life in 2030”, (2016, Sep 15th); Report of 2015 Stanford University AI study panel; No available author; Retrieved from https://ai100.stanford.edu/sites/ default/files/ai_100_report_0831fnl.pdf ;
“Artificial Intelligence, Robotics, Privacy and Data protection “; (2016,Oct); Document from 38th International conference of Data Protection and Privacy Commissioners; No available author; Retrieved from https://secure.edps.europa.eu/EDPSWEB/ webdav/site/mySite/shared/Documents/ Cooperation/Conference_int/ 16-10-19_Marrakesh_AI_paper_EN.pdf;
“Mountain Gorilla”; (retrieved on Dec 23rd, 2016); Wikipedia site; Retrieved from https://en.wikipedia.org/wiki/Mountain_gorilla ;
“Mountain Gorillas : Close relatives at risk”; (retrieved on Dec 23rd,2016); World Wide Fund for Nature website; Retrieved from https://www.wwf.org.uk/wildlife/mountain-gorillas ;
“Protection Assistant for Wildlife security publications”; (retrieved on Dec 30th, 2016); PAWS website; Retrieved from http://teamcore.usc.edu/people/Paws/index.html ;
Brown, E. (2012, Mar 7th); “Gorilla DNA offers clues about humans too”; Retrieved from http://articles.latimes.com/2012/mar/07/science/la-sci-gorilla-genome-20120308 ;
Heimbuch, J. (2010, Sep 21st); “New software helps endangered zebras by analyzing people”; Retrieved from http://www.treehugger.com/clean-technology/new-software-saves-endangered-zebras-by-analyzing-people.html ;
Kratochwill, L. (2016, Apr 22nd);” Artificial Intelligence Fights Wildlife Poaching”; Popular Science; Retrieved from http://www.popsci.com/national-science-foundation-fights-poaching-with-artificial-intelligence ;
Rutagarama, E. (retrieved on Jan 10th, 2016); “Closer Look : International Gorilla Conservation Programme”; Retrieved from http://www.fauna-flora.org/closerlook/international-gorilla-conservation-programme/ ;
Shah, P. (2012, Dec 22nd); “How do you explain Machine Learning and Data Mining to non-computer science people”; Retrieved from https://www.quora.com/How-do-you-explain-Machine-Learning-and-Data-Mining-to-non-Computer-Science-people ;
Snow, J. (2016, Jun 16th); “Rangers use Artificial Intelligence to Fight Poachers”; National Geographic;Retrieved from http://news.nationalgeographic.com/2016/06/paws-artificial-intelligence-fights-poaching-ranger-patrols-wildlife-conservation/ ;
Verayo, A. (2016, Jan 06th); “Crowd Sourcing and Artificial Intelligence Could Solve World’s Problems”; Retrieved from http://en.yibada.com/articles/99397/20160106/crowd-sourcing-artificial-intelligence-solve-worlds-problems.htm .