September 25th - 26th | Online

Bioacoustics & AI 101

Making AI accessible to everybody

Organised by

Concept

Recent developments in AI have not only led to dramatic increase in accuracy of detecting/classifying sounds, but have simultaneously made these tools accessible for people with little to no prior knowledge of AI. This workshop aims at a demystifying AI and highlighting the newest innovative research on bioacoustics machine learning tools and possible paths for the future.

This workshop is ideal for biologists and ecologists curious about AI applications, researchers seeking new sound analysis techniques, and anyone interested in the power of AI for environmental monitoring.

Objectives

Demystifying AI

We will break down complex AI concepts into easy-to-understand language, making it accessible for everyone, regardless of prior AI knowledge.

Hands-on Learning

This workshop will guide participants through training their own deep learning network to classify sounds without writing any code.

Cutting-Edge Research

We will explore the latest research in AI-powered bioacoustics, showcasing its potential and future directions.

Information

Abstract Deadline

2nd of August 2024

Acceptance Notification

23rd of August 2024

Registration Closes

13th of September 2024

Enroll Now

Latest Research

Discover the Latest Research of Our Speakers

What determines the information update rate in echolocating bats

This study investigates how four different bat species adjust their inter-pulse intervals (IPI) during commute flights, revealing that the sensory update rate varies with changes in echolocation signal characteristics and flight parameters.

BovineTalk: machine learning for vocalization analysis of dairy cattle under the negative affective state of isolation

This project aims to enhance the understanding of cattle communication and improve welfare assessments by utilizing advanced computational techniques to classify vocalizations into high and low frequency, as well as identify individual cows based on their vocal characteristics.

ReFrogID: Pattern Recognition for Pool Frog Identification Using Deep Learning and Feature Matching

This research introduces ReFrogID, a deep learning-based method that identifies individual frogs using their unique abdominal patterns, demonstrating important segmentation and pattern matching performance. This technique not only aids in monitoring endangered species effectively but also showcases the application of advanced machine learning techniques in ecological research.

Using deep learning to track time × frequency whistle contours of toothed whales without human-annotated training data

This article explores deep learning to track toothed whale whistles without human annotations. It introduces a method using automatically generated pseudo-labels, improving whistle extraction performance. This efficient approach aids in species classification and identification in marine mammal research.

Acoustic connectivity of alpine valley ecological corridors : A study on soundscapes and biodiversity protection

This project focused on the acoustic connectivity of soundscapes in alpine valley ecological corridors aimed at protecting alpine biodiversity and ecosystems. The findings demonstrated a notable difference in soundscapes due to habitat types and proximity to valley centers, which are more anthropized and contribute to corridor fragmentation.

Vocal Behavior in Spotted Seals (Phoca largha) and Implications for Passive Acoustic Monitoring

This research discusses the vocal behavior of spotted seals (Phoca largha) and their potential for passive acoustic monitoring. The study outlines the vocal repertoire of the seals, emphasizing underwater vocalizations essential for communication and behavior analysis.