New tools track food nets and Sound Skips

Artificial intelligence is opening new land in the environment. At the University of Rice, Kesar A andib are developing computational tools to help scientists better understand the environmental system with recent studies, which are used to achieve new insights from a variety of environmental data.

“AI allows us to analyze environmental data in ways that were not possible before,” said Louis Owen’s Assistant Professor of Electrical and Computer Engineering and a member of the Ken Kennedy Institute in Rice.

“These recent projects look at two different questions using different types of data from two different continents. We can spread a huge set of regions with these tools and spread to the types of data.”

A project introduces a new way to compare biological networks. The net of interaction between species identifying each ecosystem. The purpose is to identify structural similarities between the ecosystem in different areas, even when they are made of completely different species. Such a comparison can guide the environmental system’s widespread monitoring and protection priorities. Traditional ways, however, often struggle with this complex data.

Orb, in combination with Ladia Bidroot and other colleagues at Michigan State University, apply a new section of mathematics, known as more and more transporting distances to analyze the food nets of more than a hundred African eaten foods in six different regions of the continent.

Are published in studies Ways in environment and evolution.

Most transport describes the minimum amount of work needed to convert one thing to another. If each item is represented by a hill of dirt, then the maximum transportation, or the distance of the “Earth Moor”, represents the most effective way to move the dirt around so that the two mounds become uniform. In the environment, every network of species interaction can be considered as one of them as a “mound”.

Maximum transport distance allows researchers to align the overall structure of two networks, which shows how their contact samples are compared even when the network consists of different species.

Using these tools, researchers analyzed data from several sources and practically managed to identify equal species, that is, different species that play the same environmental role in their own ecosystem.

“This allows us to determine, for example, if in this food web, the lion plays a role like a leopard in it or in the other,” said Yurib.

Attempts to correct the amount of environmental data were led by former Rice undergraduate Kai Hang, who is now a doctorate at the Massachusetts Institute of Technology, and Alex Zillus, who is now pursuing a doctorate at the University of California, Berkeley.

Orib said, “He did so well in rice here that he was recruited into the nation’s high programs.” “It really talks about training and undergraduate research experience we provide.”

Earlier, a project was focused on Colombia’s tropical forests and used sound to map biological diversity. In a Colombia, headed by a doctorate student, Maria Gorroo, who is joining rice this fall as a scholar on Fulbright Scholarship, was installed in a range of residences within Colombia’s oil palm gardens. For more than 10 days, the team recorded hundreds of hours of sound, in which frogs, birds and insects received calls.

Through the AI ​​analysis, the researchers created the same what Yurib called “contacting the tropical forests”, borrowing a term from neuro science to explain how different parts of the jungle are linked by sound.

“Instead of contacts inside the brain, we were watching the contacts in the tropical forest – how information and energy flow,” said Yurib.

“We were using biocostics data as a proxy to understand the health status of an ecosystem. The novelty here is able to automatically identify and separate the sounds.”

The results show that more than the residence makes more than distances: two patches of the intact forest may take the same, even far away, while a forest and nearby region that are applied to oil palms can be quite different. This study confirms the fact that converting local forests into collective gardens rapidly reduces biological diversity, which highlights how biocostics can act as a low -cost tool for large -scale monitoring.

For Colombia, and forbia, the project had a special weight.

Orib said, “This is personally meaningful because I’m researching which has a global impact, using techniques that I am developing here with many local, regional and international colleagues in the United States.”

“In terms of effect, both papers are meaningful because research applies to AI for anything other than earning maximum profit or gaining competitive superiority: it is AI for environment and protection.”