PaperSwipe

State Space Models for Bioacoustics: A comparative Evaluation with Transformers

Published 3 days agoVersion 1arXiv:2512.03563

Authors

Chengyu Tang, Sanjeev Baskiyar

Categories

cs.SDcs.AI

Abstract

In this study, we evaluate the efficacy of the Mamba model in the field of bioacoustics. We first pretrain a Mamba-based audio large language model (LLM) on a large corpus of audio data using self-supervised learning. We fine-tune and evaluate BioMamba on the BEANS benchmark, a collection of diverse bioacoustic tasks including classification and detection, and compare its performance and efficiency with multiple baseline models, including AVES, a state-of-the-art Transformer-based model. The results show that BioMamba achieves comparable performance with AVES while consumption significantly less VRAM, demonstrating its potential in this domain.

State Space Models for Bioacoustics: A comparative Evaluation with Transformers

3 days ago
v1
2 authors

Categories

cs.SDcs.AI

Abstract

In this study, we evaluate the efficacy of the Mamba model in the field of bioacoustics. We first pretrain a Mamba-based audio large language model (LLM) on a large corpus of audio data using self-supervised learning. We fine-tune and evaluate BioMamba on the BEANS benchmark, a collection of diverse bioacoustic tasks including classification and detection, and compare its performance and efficiency with multiple baseline models, including AVES, a state-of-the-art Transformer-based model. The results show that BioMamba achieves comparable performance with AVES while consumption significantly less VRAM, demonstrating its potential in this domain.

Authors

Chengyu Tang, Sanjeev Baskiyar

arXiv ID: 2512.03563
Published Dec 3, 2025

Click to preview the PDF directly in your browser