We present a method to develop low-complexity convolutional neural networks (CNNs) for acoustic scene classification (ASC). The large size and high computational complexity of typical CNNs is a bottleneck for their deployment on resource-constrained devices. We propose a passive filter pruning framework, where a few convolutional filters from the CNNs are eliminated to yield compressed CNNs. Our hypothesis is that similar filters produce similar responses and give redundant information allowing such filters to be eliminated from the network. To identify similar filters, a cosine distance based greedy algorithm is proposed. A fine-tuning process is then performed to regain much of the performance lost due to filter elimination. To perform efficient fine-tuning, we analyze how the performance varies as the number of fine-tuning training examples changes. An experimental evaluation of the proposed framework is performed on the publicly available DCASE 2021 Task 1A baseline network trained for ASC. The proposed method is simple, reduces computations per inference by 27%, with 25% fewer parameters, with less than 1% drop in accuracy.
Data-Efficient Low-Complexity Acoustic Scene Classification via Distilling and Progressive Pruning
Shanghai Jiao Tong University（上海交通大学）；Tsinghua University（清华大学）；North China Electric Power University（华北电力大学）；Huakong AI Plus Company Limited（华科人工智能有限公司）
Switchable deep beamformer for high-quality and real-time passive acoustic mapping
School of Information Science and Technology, ShanghaiTech University（上海科技大学信息科学与技术学院）；Gene Editing Center, School of Life Science and Technology, ShanghaiTech University（上海科技大学生命科学与技术学院基因编辑中心）；State Key Laboratory of Advanced Medical Materials and Devices, ShanghaiTech University（上海科技大学先进医用材料与设备国家重点实验室）；Shanghai Engineering Research Center of Intelligent Vision and Imaging, ShanghaiTech University（上海科技大学智能视觉与影像上海工程研究中心）；Key Laboratory of Biomedical Imaging Science and System, Chinese Academy of Sciences（中国科学院生物医学成像科学与系统重点实验室）
AnuraSet: A dataset for benchmarking Neotropical anuran calls identification in passive acoustic monitoring
Instituto de Investigación de Recursos Biológicos Alexander von Humboldt（亚历山大·冯·洪堡生物资源研究所）；Cornell University（康奈尔大学）；Pontificia Universidad Javeriana Seccional Cali（哈维里亚那天主教大学卡利分校）；Universidade Estadual de Campinas（坎皮纳斯州立大学）；Museu de Diversidade Biológica (MDBio), Universidade Estadual de Campinas（坎皮纳斯州立大学生物多样性博物馆）；Instituto de Pesquisa da Biodiversidade (IPBio)（生物多样性研究所）；Universidade Federal de Mato Grosso do Sul（南马托格罗索联邦大学）；Universidade Federal de Santa Catarina（圣卡塔琳娜联邦大学）；Universidade Federal de Goiás（戈亚斯联邦大学）；Universidad Autónoma de Madrid（马德里自治大学）；Centro de Investigación en Biodiversidad y Cambio Global (CIBC), Universidad Autónoma de Madrid（马德里自治大学生物多样性与全球变化研究中心）；Laboratório de Herpetologia e Comportamento Animal, Departamento de Ecologia, Instituto de Ciências Biológicas, Universidade Federal de Goiás（戈亚斯联邦大学生物科学学院动物行为与爬行动物学实验室）
Passive acoustic non-line-of-sight localization without a relay surface
Department of Applied Physics, Hebrew University of Jerusalem（应用物理系，耶路撒冷希伯来大学）；Alexander Grass Center for Bioengineering, Hebrew University of Jerusalem（亚历山大·格拉斯生物工程中心，耶路撒冷希伯来大学）
SEABAD: A Tropical Bird Activity Detection Dataset for Passive Acoustic Monitoring
Faculty of Computer Science and Information Technology, Universiti Malaya（马来亚大学计算机科学与信息技术学院）；Faculty of Electrical Engineering, Universiti Teknologi Malaysia（马来西亚科技大学电气工程学院）
A Lottery Ticket Hypothesis Framework for Low-Complexity Device-Robust Neural Acoustic Scene Classification
Georgia Institute of Technology（佐治亚理工学院）；University of Enna Kore（恩纳 Kore 大学）；University of Science and Technology of China（中国科学技术大学）；Tencent Media Lab, Tencent Corporation（腾讯公司腾讯媒体实验室）
