• Home
  • Search
  • Broad learning system based on attention mechanism and tracking differentiator
  • https://doi.org/10.3724/sp.j.1249.2024.05583Copy DOI Icon

Broad learning system based on attention mechanism and tracking differentiator

Show More
  • Abstract
  • Literature Map
  • Similar Papers
Abstract

Broad learning system (BLS) has advantages such as a simple model structure, high training efficiency, and easy interpretability. However, it also has drawbacks such as insufficient feature learning capability and unstable generalization performance. To alleviate these problems, broad learning system based on attention mechanism and tracking differentiator (TD), abbreviated as A-TD-BLS, was proposed. In terms of model structure, A-TD-BLS introduced self-attention mechanism to the original BLS, and further fused and transformed the extracted features through attention weighting to improve the feature learning ability.In terms of model training methods, a weight optimization algorithm based on tracking differentiator was designed.This method effectively alleviates the overfitting phenomenon of the original BLS by limiting the size of the weight values, significantly reduces the influence of the number of hidden layer nodes on model performance and makes the generalization performance more stable.Moreover, the training algorithm was extended to the BLS incremental learning framework, so that the model can improve performance by dynamically adding hidden layer nodes.Multiple experiments conducted on some benchmark datasets show that compared to the original BLS, the classification accuracy of A-TD-BLS is increased by 1.27% on average on classification datasets and the root mean square error of A-TD-BLS is reduced by 0.53 on average on regression datasets.Besides, A-TD-BLS is less affected by the number of hidden layer nodes and has more stable generalization performance. Based on the above experimental results, it can be concluded that A-TD-BLS enhances the stability of generalization performance of the original BLS model, reduces the sensitivity of the model's generalization performance to hyperparameters, and effectively suppresses the phenomenon of overfitting.

Similar Papers
  • Research Article

Rapid Detection of Adulteration in Birch Sap Using Near-Infrared Spectroscopy Coupled With Chemometrics.

  • Apr 01, 2026
  • Journal of food science
  • Pengtao Zhang +6
  • PDF
  • Research Article
  • Citations8

Network Intrusion Detection Based on Deep Belief Network Broad Equalization Learning System

  • Jul 31, 2024
  • Electronics
  • Miaolei Deng +5
  • Conference Article
  • Citations1

Positive Style Accumulation: A Style Screening and Continuous Utilization Framework for Federated DG-ReID

  • Oct 27, 2025
  • Xin Xu +6
  • Research Article
  • Citations13

From WASD to BLS with application to pattern classification

  • Apr 30, 2021
  • Applied Soft Computing
  • Mei Liu +4
  • Conference Article

Layer-Wise Unlearning for Model Adaption in Non-Stationary Environments

  • Nov 12, 2025
  • Yanbing Zhou +5
  • Research Article
  • Citations1

A System of Multimodal Image‐Text Retrieval Based on Pre‐Trained Models Fusion

  • Dec 10, 2024
  • Concurrency and Computation: Practice and Experience
  • Qiang Li +8
  • Research Article
  • Citations19

Investigation on lightweight identification method for pavement cracks

  • Sep 02, 2024
  • Construction and Building Materials
  • Anxin Meng +6
  • Conference Article
  • Citations3

Dense Broad Learning System with Proportional Integral Differential and Adaptive Moment Estimation

  • Dec 01, 2020
  • Weidong Zou +2
  • Research Article
  • Citations2

Enhanced Emotion Recognition in Videos: A Convolutional Neural Network Strategy for Human Facial Expression Detection and Classification

  • Mar 24, 2023
  • Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
  • Arselan Ashraf +5
  • Research Article
  • Citations8

A Robust Strategy to Account for Data Sampling Variability in the Development of Hydrological Models

  • Mar 01, 2023
  • Water Resources Research
  • Feifei Zheng +5
  • Research Article

A multi-stage training and deep supervision based segmentation approach for 3D abdominal multi-organ segmentation.

  • Jul 17, 2025
  • Journal of X-ray science and technology
  • Panpan Wu +7
  • Research Article
  • Citations1

A Salient Object Detection Network Enhanced by Nonlinear Spiking Neural Systems and Transformer.

  • Jun 20, 2025
  • International journal of neural systems
  • Wang Li +4
  • Research Article
  • Citations29

Influence of data quality on the performance of supervised classification models for predicting gravelly soil liquefaction

  • Aug 04, 2023
  • Engineering Geology
  • Jilei Hu +1
  • Research Article
  • Citations62

Multi-Attn BLS: Multi-head attention mechanism with broad learning system for chaotic time series prediction

  • Nov 21, 2022
  • Applied Soft Computing
  • Liyun Su +2
  • Research Article
  • Citations3

How to Collaborate: Towards Maximizing the Generalization Performance in Cross-Silo Federated Learning

  • Apr 01, 2025
  • IEEE Transactions on Mobile Computing
  • Yuchang Sun +2
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.