It is not the first time, nor will it be the last, that in this blog we talk about how artificial intelligence (AI) can help preserve animals, nature, prevent earthquakes, identify diseases, improve people’s lives with autism, and even who knows if one day we will be able to communicate with our favorite pets, dogs. But today we are going to talk about Wild Me, a non-profit organization that has developed Wildbook, a data management platform and a framework for wildlife research and conservation using AI.
And what was the origin of this project? Jason Holmerg, Director of Engineering at Wild Me and Head of Information Architecture at Wildbook, joined a scientific expedition at the end of 2002, and learned that plastic tags would be thrown on whale sharks to track their movements and obtain a more accurate count of the population, thus better understanding its place in the marine ecosystem. However, the marks left by the tags on sharks were rarely seen again over time – less than one percent of the time – meaning that this method of animal identification was not very effective.
Holmberg believed that the identification rate could be improved in a different way, namely by developing a computer vision algorithm that recognized the unique spotting pattern of each shark. But then he encountered an unexpected difficulty in being able to share this tool with researchers around the world: “ Obviously we needed a web-based platform so that everyone could access it through their browser. At that time there was no such platform, nor was there anything that came close to it, so I had to devise it from scratch . “
There are still many species of living beings to know
This is what prompted the formation of Wild Me, now chosen by Microsoft as one of its most recent projects within AI for Earth, to provide new resources and support with the aim of expanding and accelerating this scientific conservation work. Several independent analyzes show that species are going extinct at an unprecedented rate around the world. In March 2018, the Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES) published a set of reports that concluded that animal and plant species are severely threatened in almost all regions of the world.
Scientists have observed and described only 1.5 million species of the 10 million estimated on Earth. And to advance these challenges, researchers need to collect, process, and analyze more data on species more efficiently than ever before, so they can strategize and win public support. However, many research teams lack the tools and resources to collect that data and work together on a large scale.
And it is at this point that Wild Me and the Wildbook project, which combines a wide diversity of functions and disciplines in a common cloud-based platform, have their full meaning. The tools that cloud services provide for data collection, storage, analysis, and collaboration help researchers work at scale.
Thanks to the availability and widespread use of digital cameras, along with innovations in storage technology and automatic image analysis software, there is a large number of snapshots that are a very inexpensive source of data. This, in turn, can be used to produce a database that records the identity, location, time, and date that the animals were photographed.
And with the computer vision algorithms implemented by Wild Me, individuals of a growing number of species can be identified by their unique patterns of spots, stripes, or other markings and physical characteristics. Combining this with geographic, environmental, behavioral, and climatic data helps answer ecological and conservation questions, such as population sizes, species interactions, and movement patterns.
“The world is facing a major biodiversity crisis, and Wild Me’s work on harnessing vision and machine learning to monitor and track individual animals is truly groundbreaking,” says Bonnie Lei, project manager at IA for Earth. from Microsoft. But going forward, Wild Me and Microsoft have more joint plans.
Sharks on YouTube
One of these initiatives is a smart agent for Whale Sharks. Their job is to check YouTube every night for videos titled or tagged “whale shark.” This agent uses machine learning to determine if the video has relevant material of a whale shark; and Microsoft Translator, with natural language processing, to read the video description and determine when and where the shark was seen. The agent can also use the same natural language and translation tools to automatically request from the video author, in their own language, information about when and where the video was taken in case it cannot be determined from the post itself.
Wild Me is also building a Twitter bot called Tweet-A-Whale, which could interact with people on whale watching tours. In this way, people will be able to Tweet pictures of whales with a time and place for the bot to automatically identify it, taking advantage of the potential of AI and machine learning to individually recognize each of the whales that appear in the photos .
Exciting, you can not say otherwise, the projects that Wild Me and Microsoft have in hand, to continue helping to recover the Biodiversity of the Earth. We will continue to count these types of initiatives, which clearly demonstrate the benefits of the application of artificial intelligence for the benefit of the Earth and those who inhabit it.
