Online Domain-Incremental Learning Approach to Classify Acoustic Scenes in All Locations

被引:0
|
作者
Mulimani, Manjunath [1 ]
Mesaros, Annamaria [1 ]
机构
[1] Tampere Univ, Signal Proc Res Ctr, Tampere, Finland
来源
32ND EUROPEAN SIGNAL PROCESSING CONFERENCE, EUSIPCO 2024 | 2024年
关键词
Domain-incremental learning; online learning; acoustic scene classification; Batch Normalization layers; forgetting; deep learning model; NEURAL-NETWORKS;
D O I
10.23919/EUSIPCO63174.2024.10715156
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
In this paper, we propose a method for online domain-incremental learning of acoustic scene classification from a sequence of different locations. Simply training a deep learning model on a sequence of different locations leads to forgetting of previously learned knowledge. In this work, we only correct the statistics of the Batch Normalization layers of a model using a few samples to learn the acoustic scenes from a new location without any excessive training. Experiments are performed on acoustic scenes from 11 different locations, with an initial task containing acoustic scenes from 6 locations and the remaining 5 incremental tasks each representing the acoustic scenes from a different location. The proposed approach outperforms fine-tuning based methods and achieves an average accuracy of 48.8% after learning the last task in sequence without forgetting acoustic scenes from the previously learned locations.
引用
收藏
页码:96 / 100
页数:5
相关论文
共 36 条
  • [31] Machine learning the gap between real and simulated nebulae: A domain-adaptation approach to classify ionised nebulae in nearby galaxies
    Belfiore, Francesco
    Ginolfi, Michele
    Blanc, Guillermo
    Boquien, Mederic
    Chevance, Melanie
    Congiu, Enrico
    Glover, Simon C. O.
    Groves, Brent
    Klessen, Ralf S.
    Mendez-Delgado, J. Eduardo
    Williams, Thomas G.
    ASTRONOMY & ASTROPHYSICS, 2025, 694
  • [32] Not All Data Matters: An Efficient Approach to Multi-Domain Learning in Vehicle Re-identification
    Quang Truong
    Mei, Bo
    2021 IEEE INTELLIGENT TRANSPORTATION SYSTEMS CONFERENCE (ITSC), 2021, : 1278 - 1285
  • [33] Augmenting fake content detection in online platforms: A domain adaptive transfer learning via adversarial training approach
    Ng, Ka Chung
    Ke, Ping Fan
    So, Mike K. P.
    Tam, Kar Yan
    PRODUCTION AND OPERATIONS MANAGEMENT, 2023, 32 (07) : 2101 - 2122
  • [34] Forget less, count better: a domain-incremental self-distillation learning benchmark for lifelong crowd counting忘得少, 数得好: 一种域增量式自蒸馏终身人群计数基准
    Jiaqi Gao
    Jingqi Li
    Hongming Shan
    Yanyun Qu
    James Z. Wang
    Fei-Yue Wang
    Junping Zhang
    Frontiers of Information Technology & Electronic Engineering, 2023, 24 : 187 - 202
  • [35] A novel approach of online monitoring for laser powder bed fusion defects: Air-borne Acoustic Emission and Deep Transfer Learning
    Li, Zhiwen
    Zhang, Zhifen
    Zhang, Shuai
    Bai, Zijian
    Qin, Rui
    Huang, Jing
    Wang, Jie
    Huang, Ke
    Zhang, Qi
    Wen, Guangrui
    JOURNAL OF MANUFACTURING PROCESSES, 2023, 102 : 579 - 592
  • [36] A transfer learning approach for acoustic emission zonal localization on steel plate-like structure using numerical simulation and unsupervised domain adaptation
    Ai, Li
    Zhang, Bin
    Ziehl, Paul
    MECHANICAL SYSTEMS AND SIGNAL PROCESSING, 2023, 192