How the Wildlife AI Platform Solves Data Challenges – Illinois News Today

Artificial Intelligence
Image credit: source

Wild Me is a non-profit machine learning service provider for field biologists studying wildlife and conservation. But before we can create a whale shark algorithm, we need the right data.

Anyone involved in data management and data science can prove the challenge and time-consuming nature of mapping datasets from new sources to the platform. The platform can clean up, validate, and finally analyze the data for use in algorithm training. After all, your algorithms are just as good as the data used to train them.

Now, whether these datasets are from hundreds of external users who are using any number of systems to collect this data, from Excel files to real shoeboxes full of photos. Imagine. This is a challenge for non-profit wildlife conservation machine learning and artificial intelligence service providers. Wild Me We have been facing operations for over 10 years. The organization is building open software and AI for the Nature Maintenance Research Community. The organization is made up of engineers who are professionals in software and machine learning and is designed to be “a reliable engineering driving force for wildlife biologists around the world.”

With this AI software, researchers can track individuals of different species (such as whale sharks) and identify them with unique spot patterns. Wild Me created this first use case algorithm and technology through a modification of the Hubble Space Telescope algorithm that examines the patterns of stars in the night sky. Jason Holmberg, Executive Director, Co-Founder, and Engineering Director of the organization.

Jason Holmberg Credits: Via WildMe

During his 2002 scuba trip in Djibouti, he saw the first whale shark and learned how researchers could physically tag and track animals. He thought there might be a better way through computer vision algorithms that could identify individuals by their own spot patterns. This work has become Whaleshark.org, a library of encounters and individual whale sharks used and maintained by marine biologists.

But that was just the first use case. From there, WildMe has been extended as a platform for other animal researchers, allowing data to be uploaded and cataloged from a range of other species. From Manta Ray to Giraffe and Sea Dragon.. The platform serves more than 200 organizations and nearly 1,000 researchers, has nearly 444,000 sightings in its database, and tracks nearly 90,000 animals worldwide.

The challenge of moving biologist encounters, sightings, and personal catalogs to the Wild Me platform has been a daunting task from the beginning.

“It was an evolving process,” Holmberg said. “When I first started working with biologists around the world, I was creating a custom importer for all my data. That custom one-off code took weeks.”

Ben Shiner, Wild Me’s senior software engineer explains: “I had my own hand-wound JavaScript framework for importing data, but there was a bug. It focuses on environmental issues and AI and machine learning are important services. Understand this data onboarding Deserves its own company and set of solutions, something that couldn’t be done with a non-profit bank account. “

Ben Shiner Credit: Via WildMe

Ben Shiner Credit: Via WildMe

There was no universal standard for how individual researchers could catalog their data. Each researcher has created his own system.

For this reason, “The idea of ​​a universal data importer is a kind of farce,” Holmberg said. “But we were able to solve half of the problem.” Wild Me began using tools to help field biologists begin mapping data to a common set of fields and descriptors. These biologists can review and approve the data in the system.

This streamlined and speeded up the process, but there were still issues that could be improved. The system wasn’t very scalable, and researchers couldn’t even validate their data. Wild Me has started trial operation of a company tool called FlatfileIt is designed to solve the problem of processing and validating external data from multiple sources.

David Boscovic After working for several different SaaS companies and encountering the same nasty issues each time, he founded Flatfile. That is, when each customer uses a different system, the new customer’s data is brought into the system.

“This was a universal issue. The cost and effort of capturing data is one of the costs of innovation,” says Boskovic. But it was very frustrating. “I would say I designed this product with rage.”

David Boskovic Credit: Via Flatfile

David Boskovic Credit: Via Flatfile

Another aspect of getting data into the system is that customers need to maintain ownership and control of that data. It’s important for marketers. It is also important for field biologists. This is one of the reasons WildMe pursued pilots in Flatfiles.

“It’s an intuitive system, where field biologists can maintain ownership of the data through the process of importing it into the system and do things like data validation that we didn’t currently have.” Holmberg said. For example, “Make sure all GPS coordinates are in the correct format. These are human-created data catalogs. There are errors.”

During the validation process, the biologist who curated the data will be presented with anomalies and will be able to go back and clean up the data. This allows biologists to view and manipulate data on one of the Wild Me platforms.

The platform is changing the knowledge of species that biologists study.

“When I first started studying whale sharks, everyone thought the Indian Ocean was a big place for that,” Holmberg said. “When we built these online platforms, we were able to identify individual behavior … the Gulf of Mexico now turns out to be one of the biggest hotspots for studying whale shark behavior. It was. “

Wild Me is often the first experience of researchers in cloud computing, data storage and analysis, so the goal is to make the system easier to use for non-technology-focused people.

Holmberg said data processing needs to be fast so that biologists can adapt to changing populations with better conservation policies and strategies.

“It may mean raising or lowering fences, allowing fishing, or banning fishing, depending on the impact of variables on the population,” he said. “The faster the population is estimated, the faster we will respond to change, and we will conserve for more successful solutions, especially to help increase the population of endangered animals. You can see that the strategy is repeating. “

What to read next:

From AI to Teamwork: Seven Key Skills for Data Scientists
The basics of machine learning that everyone should know
How to hire AI talent and keep them happy
Become a self-taught cybersecurity professional

Jessica Davis is a senior editor at InformationWeek. She covers enterprise IT leadership, careers, artificial intelligence, data and analytics, and enterprise software. She spent a career covering the crossroads of business and technology. Follow her on Twitter: … View full biography

Comments on this topic are welcome on social media channels. [contact us directly] With questions about the site.

Other insights

(Excerpt) Read more Here | 2021-07-03 18:30:49

Leave a Reply

Your email address will not be published. Required fields are marked *