Artificial intelligence learning to listen to nature

Acoustic monitoring is a common technique to survey threatened animal species. By deploying small recording devices in the field, researchers record sounds which can be played back later to identify which species are calling. While this method makes it incredibly easy to gather large amounts of data, these massive datasets of thousands of hours of audio then need to be reviewed manually.

Researchers from the Arthur Rylah Institute (ARI) have developed innovative sound-recognition software to identify some of Victoria’s frogs. This utilises technology called 'deep learning' which is a type of Artificial Intelligence (AI) that teaches computers to process data in a way inspired by the human brain. It has the potential to save scientists thousands of hours by listening to frog recordings on their behalf.

The software can be trained to listen to recordings and identify species. This dramatically increases data processing accuracy and efficiency. The software uses technology like those used in image or song recognition programs. Currently, the software can analyse up to 1 hour of recordings every 10 seconds.  

A unique feature of this software is the ‘find it’ tool which helps with call validation and matching. This automatically directs researchers to specific points in the recording so they can check the program has correctly identified a call or listen to any unusual sounds. For example, frogs detected outside their usual location. 

Using the software has saved staff significant time, which they otherwise would have spent on sample analysis. In 2019-20, the Living Murray Frog monitoring project collected around 3,225 hours of frog recordings from Barmah-Millewa Forest. With traditional methods, not using the software, a staff member took 3 weeks to process only 10% of these calls. In 2020-21 the software took 3 days to analyse approximately 4,835 hours of recordings with staff only needing a few days for processing and validation.

DEECA’s Forest Protection Service Program (FPSP) and ARI funded the pilot artificial intelligence recogniser model. It was trained and tested using data collected through monitoring projects funded by The Living Murray and Regional Forestry Agreement Landscape Surveys Program. 

The software isn’t just limited to frogs either – it can be trained to listen out for many other species. ARI researchers are also using the software to listen for the threatened Eastern Bristlebird and several bat species. 

Current research on the Eastern Bristlebird is funded by the Australian Government’s Department of Agriculture Water and the Environment, FPSP and ARI’s capability fund.