Come find Tyne at the ASPB Conference! She’ll be presenting thrice!

A spectrogram with a Red bar over a harmonically rich vocalization. The caption in red text says "HUMAN".

Interested in an exciting new ML model for Bioacoustics? Check out her talk:

Find all the Humans: Developing a Machine-learned Model to Improve Human Privacy in Open-source Bioacoustic Datasets

Wednesday, November 22 at 3:05pm

Well labelled open source datasets serve as the basis for creating software tools for biologists, and furthering large-scale research efforts. When recording in open or public spaces, human voices are often captured incidentally. There are currently few available or effective options for removing voices in large-scale datasets. We’re are developing a brand new Machine-learned model and associated application to help researchers protect the public’s privacy while open-sourcing more large labelled ecoacoustic datasets. This talk and model are being developed in collaboration with skilled Software Developer Jarrett Lubky. Come hear about our progress on this exciting new project!



Photo of an angry marmot. There is a red detection box on the image with white letters reading "animal".

How are we feeling about Machine Vision? AI for ecology? Large Audio-visual dataset management? Lost, fearful, excited, or all of the above? Tyne’s new workshop on Machine ID might be for you!

Overwhelming to Understanding: Machine-powered Species ID for Large Audio-visual Datasets

Friday, November 22 at 12:30-2:30pm and 3:00-5:00pm

A/Vian Eco processes large datasets regularly with computer assistance. I hear the interest, excitement, and concern around implementing these tools from our community of practitioners. If we can do it practically, so can you! Join me for one of two 2-hour workshops, where I demystify the world of machine-learned models for species identification in audio-visual datasets. Discover the historical evolution of “AutoID” in Ecology, gain high-level insights into the inner workings of machine-learned models, and explore various models currently available with practical examples. Most importantly, learn to systematically evaluate models before implementing them in your work, and know the practical validation steps for any model that will allow confident data analysis worth your professional seal.

Full conference details are available on the ASPB website:


AI Disclosure: The featured image in this post was created using Dall-e 3 and is labelled accordingly. Some coding and writing assistance from ChatGPT 3.5 and 4.0 were used for the upcoming presentations and the model to be discussed in the first talk.