Aktuelles

Unser Paper Unmute the Patch Tokens: Rethinking Probing in Multi-Label Audio Classification von Lukas Rauch, René Heinrich, Houtan Ghaffari, Lukas Miklautz, Ilyass Moummad, Bernhard Sick und Christoph Scholz ist bei ArXiv online. Abstract: Although probing frozen models has become a standard evaluation paradigm, self-supervised learning in audio defaults to fine-tuning. A key reason is that global pooling […]

Unser Paper Can Masked Autoencoders Also Listen to Birds? von Lukas Rauch, Ilyass Moummad, René Heinrich, Alexis Joly, Bernhard Sick und Christoph Scholz wurde akzeptiert bei Transactions on Machine Learning Research. Abstract: Masked Autoencoders (MAEs) learn rich semantic representations in audio classification through an efficient self-supervised reconstruction task. However, general-purpose models fail to generalize well

Unser Paper Adversarial Training Improves Generalization Under Distribution Shifts in Bioacoustics von René Heinrich, Lukas Rauch, Bernhard Sick und Christoph Scholz ist bei ArXiv online. Abstract: Adversarial training is a promising strategy for enhancing model robustness against adversarial attacks. However, its impact on generalization under substantial data distribution shifts in audio classification remains largely unexplored.

Unser Paper Foundation Models for Bioacoustics — a Comparative Review von Raphael Schwinger, Paria Vali Zadeh, Lukas Rauch, Mats Kurz, Tom Hauschild, Sam Lapp und Sven Tomforde wurde bei Ecological Informatics eingereicht. Abstract: Automated bioacoustic analysis is essential for biodiversity monitoring and conservation, requiring advanced deep learning models that can adapt to diverse bioacoustic tasks.

Das Paper Continental-scale behavioral response of birds to a total solar eclipse von David Mann, Austin Anderson, Amy Donner, Michael Hall, Stefan Kahl und Holger Klinck erscheint in der 15. Ausgabe von Scientific Reports. Abstract: Based on anecdotal evidence and research involving human observation and community science, radar analyses, and acoustical studies, birds are thought

Unsere Publikation Assessing spatial variability and efficacy of surrogate species at an ecosystem scale von Kristin Brunk, H. Kramer, M. Peery, Stefan Kahl und Connor Wood erscheint in der August-Ausgabe von Conservation Biology. Abstract: Preserving biodiversity is a central goal of conservation, but, in practice, monitoring biodi-versity often involves assessing population trends for one or

Continuous Real-Time Acoustic Monitoring of endangered bird species in Hawai‘i erscheint in der 87. Ausgabe von Ecological Informatics. Abstract: The decline of endemic bird species in Hawai‘i requires innovative conservation measures enabled by passive acoustic monitoring (PAM). This paper describes a novel real-time PAM system used in the Pōhakuloa Training Area (PTA) to reduce wildlife collisions and minimize disruptions

Unser Paper BirdSet: A Large-Scale Dataset for Audio Classification in Avian Bioacoustics von Lukas Rauch, Raphael Schwinger, Moritz Wirth, René Heinrich, Denis Huseljic, Marek Herde, Jonas Lange, Stefan Kahl, Bernhard Sick, Sven Tomforde und Christoph Scholz wurde akzeptiert bei der International Conference on Learning Representations (ICLR) 2025. Abstract: Deep learning (DL) has greatly advanced audio classification, yet

Unser Paper AudioProtoPNet: An interpretable deep learning model for bird sound classification von René Heinrich, Lukas Rauch, Bernhard Sick und Christoph Scholz erscheint in der 87. Ausgabe von Ecological Informatics. Abstract: Deep learning models have significantly advanced acoustic bird monitoring by recognizing numerous bird species based on their vocalizations. However, traditional deep learning models are